imminent mose 111 006 458 - ericsample search topic,parapsychology search topic 1. title ix and...
Post on 12-Aug-2020
2 Views
Preview:
TRANSCRIPT
ED 160 109
A UTRORTITLE
INSTITUTION
SPONS AGENCY
PUB- DATEGRANTNOTEAVAILABLE FROM
IMMINENT MOSE
111 006 458
Markey, Karen; Atherton, PaulineCNTAP: Online Training and Practice Manual for ERICData Base Searchers.Syracusa Univ.,, N.Y. EEIC Clearinghouse Crinformation Resources.National Inst. of Education (DFEW), Washington,D.C.Jun 78NIE-R-76-0027182p.; Parts of document marginally legitleSyracuse University Pr ttitg Services, 12S CollegePlace, Syracuse, New York 13210 (1E-24; $6.00 plus$0.60 postage and handling)
EDRS PRICE MF-$0.83 HC-$10.C13 Plus Postage.DESCRIPTORS *Data Bases; Imprcvement Programs; Library Education;
*Manuals; *Models; Cn Line Systems; Relevance(Information Retrieval); *Search Strategies;*Training
IDENTIFIERS DIALOG; *ERIC; Information Analysis Products;Information Specialists
ABSTRACTThis self-improvement manual developed for
experienced online searchers presents and discusses severallternative search strategies and their appropriate use tc achieve
the desired outcome. The manual is divided into two parts: (1) aneight step model of the total search process emphasizing how thedecision of the information professional affects retrieval resultsand search objectives at each step of the search process; and (2)introduction to DIALOG's ONTAP file and self - improvement exerciseswith search formulations and explanaticts selected from the questionsstored in the CNTAP file. Each exercise falls within the scope-ofinterest of one of the 16 ERIC Clearinghouses. The CNTAP file is asubset of the ERIC data base' consisting of PIE and CIJE citations forthe 1975 ERIC file and 29 simple, moderate, or difficult searchQuestions with answer sets. It was developed tc allow searchers toevaluate their output by computing recall and precision scores oftheir results with the results of a "perfect search" or answer set.The appendices include 13 search save formulations cf comacn searchfacets for "population level" and "tools" which can he accessed andexchanged between DIALOG workspaces by 'sing the procedures that areoutlined, and a comprehensive guide to DIALOG's ERIC data eases.(JFF)
********** **Reproductions
************** ** ***** *********supplied by EDRS are the best that can be made
from the orig'nal dccument.**** ********** **** 4444************************* *
,11OCA;VFN, BEEN PEPP4=1 XAC QECF,-;ED T ROL.
OWC.ANi/AT,ON47S Of QPiN,ONN
T
STZ gra DO NEB 4y QPQE-v---4 NATioNAL .NsT-ruTE
U $ DEPARTMENT OA NEALTN,EOUCATiON & WELFARENATIONAL INSTITUTE OF
EDUCATION
t=)w
ONLINE TRAINING AND PRACTICEMANUAL FOR ERIC DATA BASE
SEARCHERS
BY KAREN MARKEY AND PAULINE ATHERTON
BASED ON THE EARLIER UNPUBLISHED WORK OFCHARLES P. BOURNE, BARBARA ANDERSON,
AND JO ROBINSON
ERIC CLEARINGHOUSE ON INFORMATION RESOURCESSYRACUSE UNIVERSITY
JUNE 1978
Karen Markey is a Research Associate in the Scho)1 of InformationStudies, Syracuse University, Syracuse,_New York.
Pauline Atherton is a Professor in the School of Information Studies,Syracuse University, Syracuse, New York.
The material in this lication was prepared pursuant to a contractwith the National Institute of Education, ,United States Department ofHealth, Education, and Welfare. Contractors undertaking such projectsunder government sponsorship are encouraged to express freely their,iudgement in professional and technical matters. Prior to publication,this document was submitted to Lockheed Information Systems, Palo Alto,California and to independent online searchers for critical review anddetermination of professional competence. This publication has metsuch standards. Points of"viev or opinions, however, do not necessarilyrepresent the official view or oninions of either the reviewers or theNational Institute of Education.
Prepared under Grant No. NTE-E-76-0027.
TABLE OF CONTEPJTB
PREFACE
SECTION I: THE PROCESS OF ONLINE SEARCHING 5
Introduction to the Steps and Functions cf Online Searching
Step 1. Defining the Information Need and Search Objectives 9
Step 2. Identifying Relevant Data Bases 13
Step 3. Formulating Basic Search Logic 5® PlanningSearch Strategiea 15
Step 4. Compiling the Search Terms 29
Step 5. Ordering Output
Step 6. Conceptualizing the Search as Input to theRetrieval System 35
45
Step S. Evaluating Final Results 47
Step 7. Evaluating Preliminary Results
SECTION II: ERIC ONTAP EXERCISES FOR SELF- IMPROVEMENTOF ONLINE SEARCHING
Introduction to DIALOG's ERIC ONTAP
Sample Search Topic ,Parapsychology
Search Topic 1. Title IX and Women's SportsERIC/Educational Management
Search Topic 2. 4-1-1 Clubs
ERIC/Career Education
49
49
55
65
73
Search Topic 3. 16 Personality Factor Test
75ERIC/Counseling and Personnel Services
Search Topic 4. Primary School Reading ProgramERIC/Reading and Communication Skills 79
Search Topic 5. School Busing-
ERIC/Social Studies /Social Science Education 89
Search Topic 6. Navaho Language Textbooks
93ERIC/Languages and Linguistics
Search Topic 7. AACR RevisionERIC /Information Resources 97
Search Topic 8. Drug AbuseERIC /Teacher - Education 101
Search Topic 9. Vocational Education
ERIC/Rural Education and Small Schools 105
Search Topic 10. Piaget
ERIC/Early Childhood EducationERIC/Science, Mathematics, and Envirormiental
Education
Search Topic 11. Identification of the GiftedERIC/Hzridicapped and Gifted Children
Search Topic 12. Higher Education Enrollment ProjectionsERIC /Higher Education 123
Search Topic 13. Junior College Enrollment. ProjectionsERIC/Junior Colleges 127
Search Topic 14. Family Background of Puerto Rican ChildrenERIC/Urban Education 131
Search Topic 15. Preschool Children and Learning DisabilitiesERIC/Tests, Measurement, and Evaluation 135
APPENDIX143
Common Search Facets/Search Save Formulations 143
Search Facet 1. Primary Education 147
Search Facet 2. Elementary Education 148
Search Facet 3. Secondary Education 150
Search Facet 4. Children 152
Search Facet 5. American Indians 154
Search Facet 6. Audiovisual Aids 155
Search Facet 7. Evaluation (High Recall) 160
Search Facet 8. Evaluation (High Precision) 163
Search Facet 9. Junior College 165
Search Facet 10. Higher Education 166
Search Facet 11. Mentally, Neurologically, and EmotionallyHandicapped
Search Facet 12. Preschool
Guide to DIALOG's ERIC Data Base
168
170
171
INDEX197
PREFACE
As more and more online searchers gain experience, there appears to be a
need for online and offline aids to support searchers through the "improver"
stage of training and practice. Few examples serving this purpose can be found
in the literature, but we learned of a project in this area in early 1977. At
the suggestion of, and with the encouragement and support of Oscar Firschein at
Lockheed Information Systems, Charles Bourne and his colleagues at the Institute
for Library Research (ILR) at the University of California, Berkeley, were de-
veloping a special ERIC /DIALOG file for online training and practice. This file
was being specially designed to allow online searchers to compare their search
results with the results of a "perfect search." The Berkeley group developed
"answer sets" for some 20 search requests by exhaustively searching a file of
ERIC 1975 citations for relevant hits. The composite results of all their search
strategies for each question became the "perfect answer set." For each search
topic, they offered several representative formulations to meet different Search
objectives (e.g., high recall, high precision, or low cost).
Oscar Firschein arranged to have 1975 ERIC accessions, and the 29 search
questions with their answer sets loaded on the DIALOG system. He developed and
implemented the /EVAL command on the DIALOG system This automatically computes
a recall and recision score for any searcher's effort on the ONTAPTM
questions.
by comparing the searcher's results with the "perfect answer set."
A draft report was prepared by the ILR group to document-their work, but it
could not be published before they left In late 1977 at ERIC/IR, we de-
cided to work from their draft, with their permission, and develop an "improver's"
manual for searching the ONTO data base online.
The purpose of this publication is neither to't ain beginning searchers nor
to provide an orientation to computer -based searching of ERIC. Publications
already perform these functions quite admirably.'
Our objective has been to develop the Bourne, Robinson and Anderson draft
report into a manual that the experienced searcher can use self-improvement.
In the first sec ion a model of the total search process is discussed in order
to emphasize how the decisions of the information professional affect retrieval
results and search objectives at each step of the Trocess. The second section
begins with an i v-, a to DIALOG's ONTAP file -.which is foil wed by
"self improvement" exercises selected from the questions stored in the
ONTAP file. The f ise pro:ides an in-depth explanation of the
ONTAP file as directions from the OM? file, the EVAL capability, and selecting
self-improvement exercises are discussed in detail. Within a self-learning en-
vironment, the searcher is encouraged to analyze and compare his/her search ob-
jectives and output with that of other searchers as one proceeds with the exer-
cises.
To make this publication as useful as possible to ERIC Clearinghouse
personnel, we have identified the self-improvement exercises by ERIC Clearing-
1_Bourne, Charles P. et al. Analysis of ERIC on-Line_ File Searchin-. Proce-
dures and Guidelines for Searchil_Reort. Berkeley, CA Uni-versity of California, Institute of Library Research, and Palo Alto, CA:Lockheed Research Laboratories, Nov. 1974. (ED 101 757)
Bourne, Charles P. DIALOG Lab Workbook; trainin exercises for the Lock-heed DIALOG Information Retrieval Service. _Berkeley, CA: Universityof California, Institute of Library Research, Oct. 1976. (ED 136 751)
Brandhorst, W. T., and Sharon Jewell. Search _Strategy Tutorial; _Searcher'SKit. Washington, DC: National Institute of Education, Oct. 1973. (ED 062 763)
Timbie, Michele, and Don H. Coombs, An Interactive Information RetrievalS- rem; case studies on the use of DIALOG to search the ERIC documentfile. Stanford University, ERIC Clearinghouse on Educational Media andTechnology, Dec. 1969. (ED 034 431)
Yarborough, Judith. How to Pre are for a Corn uter Search of ERIC: a non-technical approach. 44p. Sept. 1975. (ED 110 096, $.83 miCrofiche,$2.06 photocopy, plus postage, available from EDRS).
Yarborough, Judith. "A Novice's Guide to ERIC the Data Base of Education,"Online 1, 3 (July, 1977): 24-29.
house interest. Whenever Lomputer-based demonstrations and workshops are given
or staff are boing trained for online searching, we hope this manual will be
useful. The original questions loaded into the ONTAP file did not include
questions covering every clearinghouse's scope of interest. We at ERIC/1R have
solicited search requests from five clearinghouses and will try to incorporate
the questions and "perfect answer sets" into DIALOG's
as possible.
As an aid to all ERIC online searchers, we have included in the appendix
several search saves for population level and tool concepts. These can be
accessed and exchanged between DIALOG workspaces by using the procedures ex-
plained in the introduction to the appendix. As more search saves are
developed for online searching, we will announce them n the ERIC/IR
Occasional Newsletter.
Comments about this manual are welcome and should be addressed to:
Pauline AthertonERIC/IRSyracuse UniversitySyracuse, New York 13210
SECTION ITHE PROCESS OF ONLINE SEARCHING
INTRODUCTION TO THE STEPSAND FUNCTIONS OFONLINE SEARCHING
Online searching has been divided into eight steps, each requiring
o fline activities in preparation for the online activity. As the searcher
progresses from one step to the next, he/she is faced with decisions which
affect the succeeding steps. The following outline is accompanied by an ab-
strect of the functions performed during each step in the process of online
searching. The eight steps are treated in detail In sections following
the outline:
DEFINING THE INFORMATION-NEEDAND SEARCH OBJECTIVE
Interviewing the information seeker clarifiesthe request and determines search obectives:
a) retrieve all relevant items (high recall), orb) retrieve only relevant items (high precision), orc) retrieve some relevant items (Briefsearch).
IDENTIFYING RELEVANTDATA BASES
Determining which online data base to use first,hich next, etc.
-4110=MMAINIF
5
4
FORMULATING BASIC SEARCH LOGICPLANNING SEARCH STRATEGIES
Analyzing the search topic into facets or conceptgroups. Planning approaches to search strategyfor combining concepts of the topic.
C OMPILING THE SEARCH TERMS
Choosing descriptors from the ERIC Thesaurus orother printed word lists. Selecting termsin identifier field and free text forsearching subject-conveying fields (title,abstract, etc.). Using thesaurus and alpha-betic word lists online.
6
ORDERING OUTPUT
Limiting and printing output (offline and online).Selecting an approach to search strategy whichbest satisfies the search objectives expressedby the requester.
6
CONCEPTUAL I Z I NG THE SEARCH AS INPUTTO THE RETRIEVAL SYSTEM
Arr.Insing the search terms into concepts or facets,noting most important and less important conceptroups, restricting or limiting output.
EVALUATING PRELIMINARYRESULTS
Reviewing search results, step by step, and con-sidering alternate search strategies to-meetsearch objectives (recycling steps 1-6).
EVALUATING FINALRESULTS
Determining client satisfaction with Search results.
7
STEP 1=minims.
DEFINING THE INFORMATION NEED ANDSEARCH OBJECTIVES
The first step in the search model is to identify the client's information-
need through personal or telephone into= or through written messages. Since
this activity forms the basis of the search strategy, it 1- necessary that the .-
information-need be described in-as precise, specific and complete a manner as
possible.
To ensure that the information specialist has received from the client all
that is needed for searching, it may be advisable to check whether the foil
questions have been answered:
Clarification of Re -nest
Request Negotiatio
Search Vocahulaa
SearchStrate-- Formulation
Administ ve Details
* Is there an agreement on the narrative statementof the information-need?
* Has the search objective been agreed upon?* Have the subject area terminology and litera-
ture sources been discussed?* Are there any limits or constraints to-be imposed
on output (date, language, type of informationItem)?
* Do the client and information specialist agree uponthe selection and/or exclusion of search terms?
* HaVe the options of free text and controlled termsearching been discussed?
* What vocabulary aids have been selected?
s there agreement on the formation of conceptgroups?
* Have order of search statements and combinationsconcept groups been discussed?
* What alternate search strategies have been discussed?* Have output formats and limits been decided?
at are the arrangements concerning delivery of
HI&output and billing?
* H post-hoc evaluation been arranged?
For any given search there appear to be at least three different types of
search objectives:
retrieve all relevant i e a high recall-search
2) to retrieve only relevant items - a high pre _ston search
3) to retrieve some relevant items a brief search
a) A 11a12 recall search is formulated when the information seeker needs to
find everything on the stated topic. It is necessary that the person writing a
didsertation, or conducting a- tent search finCi all the relevant citations, thus:
a high recall search is required. In this case, the searcher must include all word
variants -and- synonyms for each concept. An. example of taking into account Variant
spellings and synonyms is taken from the search topic, parapsychology, the sample
self-improvemene exercise in Section II.
TARAPSYCH? Using tl'e DIALOG truncation symbol, allows thesearcher to retriev6 variant forms of the root,e.g.., parapsychology, parapsychological, etc.,,instead of keying -in these terms separately.
SENSORY The term, extrasensory, a concept synonymous withparapsychological, is typed as a single word andas two adjacent words in order to retrieve dif-ferent spellings of the same term.
S EXTRASENSORY
Oftentimes a certain volume of nonrelevant or marginally relevant output
would have to be examined to ascertain that high recall had been obtained.
b) A highlprecision2sear'ch retrieves much relevant material, but with fewer
nonrelevant items in the output. Such a search would usually us& only specific
descriptors_o terms representing concepts in the,search with no generic'term.
searching.- AnHexample of a. high precision search formulation, the Search-topic
"Identification of the Gifted" contains two concepts which can be represented by
_-ERIC descriptors; thus, statements of a high precision search would include the
Recall is defined,as the percentage of relevant items in the file that areretrieved by the search.
Precision is defined as the percentage of retrieved -items that are relevant to2
the search topic.
descriptorst_IDENTIEICATION and GIFTED. The following example would retrieve
only ERIC accessions whose title includes both terms, GIFTED and IDENTIFICATION,
or accessions with both IDENTIFICATION and GIFTED assigned as full descriptors.
S GIFTED/Df",-TI
IDENTIFICATION/DE II
1 AND 2
Thus, in the example above, high precision is dependent upon the function of
the descriptors and the title to describe accurately the intellectual content of
the document.
c) The brief earch (called Briefsearch in this workbook) is done in response
to-the need for retrieving a few items either to lessen expenses or to perfdrm a
rapid survey of the 'file before a more comprehensive and lengthy strategy. The
Briefsearch is necessarily a low recall search, performed in such a way that max-,
imam results can be achieved with little online time. Choice of priority concepts
is a factor in designing a Briefsearch.
A good example-of the Briefsearch is in Search- Topic #6, the Revision of the
Anglo-American Cataloging Rules. In a single statement, using DIALOG's word prox-
, feature, the search is performed;
S ANGLO (1W)CATALOGING(C)REVISION The use of word, proximity "(1W)"accounts for spelling variationsand the use of the full text search-ing capability (C) eliminates thenecessity of entering additionalSELECT tnd gommINE operations.
More detailed formulations of the topic are given in Section II; as you proceed
through the self-improvement exercises, compare the performances of the Briefsearch
with the other formulations in terms of recall, precision, did online connect time.
In most.cases, the search queries in Section II contain sample formulations for each
f these' three types of search obJectives.
STEP 2
IDENTIFYINGRELEVANT DATA BASES
h over 150 bibliographic databases online at this writing, this step
in the search process is becoming more and more important. For the purposes
of)the manual, however, it is assumed that ERIC is the first choice and we will--just move on.
For help with this step, the reader is referred to such articles as:
Donati, Robert. "Selective Survey of Online Access to Social Science 'Data_Bases." Special Libraries 68, Iljislov. 1977): 396-406.
Donati, Robert. "Survey of Online'Access to Social Science Data Bases."'Paper presented at the Special Libraries Association Annual Conference,Denver, CO, June 8, 1976. 23p. (ED 128 011, $.83 microfiche, $1.67"photocopy, plus postage, available from EDRS).
Hawkins, Donald T. "Multiple Database Searching:- techniques and pitfalls."Online 2, 2 (April 1978): 9-15.
Teitelbaum, Henry H., and Donald T. Hawkins. "Database Subject index."Online 2, 2 (April 1978): 16-21.
Wenger, Judith. "Multiple Database Use.' Online 1, 4 (Oct. 1977) 35-41.
STEP 3wt.mrmia
FORMULATING BASICSEARCH, LOGIC
PLANNING SEARCH STRATEGIES
-There is no single way to formulate a subject search, nor one best way given
to fulfill the same search requirement. Different searchers will approach searches
n different ways, and will assemble different formulations. Some of this due
to differences in the way people approach or attack a given problem in general,.
There may be several different formulations that give essentially the same pei
formance, but all Searchers will agree that a search should be analyzed into facets
or concept groups. Facet indicates a serarate aspect of a query. For example, a
query on the effects of TV violence on children can be separated into three dis-
jinct:e neept groups (or facets, notions, parts, components):
1) Television
2) Violence--
) Children
In order-to perform an online search, these three facets are joined using
Boolean logic operators such as and, or, not. ,Since all three facets must be
psent'to satisfy the stateCinformation need, the Boolean operator "AND"-js:
used With the search terms selected to represent the three facets.
In the illustration below, .the three facets are represented by circles; the
place where all three circles intersect forms the'portion of the data base which
contains record_ with all three facets.
16
In order to retrieve citations from the intersection of the three concept,
groups, each facet is represented by any number. of synonymous terms. These terms
are then asaembled into a series of search statements and combined by the Boolean
"OM operator to produce a "complete set of citationsJor that facet. Ir:the
example below, the facet VIOLENCE is made up of six search terms, crime, anti-
social behavior, aggression, emotional-response violencei and delinquency, com-
bined by the OR operator. How many of these terms and their variant forms could
be searched depends on thd searcher.
-Emotional
Delinquency
Anti Social Behavior
k
The inclusion of more synomous terms tl the set representing VIOLENCE would
relax the search requirement (or allow for more opportunity of matches) and permit
the formulatidn to be satisfied by more- citations. As a_ general principle which
is independent of .either data base or retrieval system, the inclusion of OR's in
eases the volume of output. On the other hand, the inclusion of AND's de-
creases the volume of output since every AND condition added to a formulation tends
to make the search more res riCtive4or reduce opportunities of matches) and
results in less output. Thus both the total number of output citations, and the
total number of relevant citations retrieved tend to decrease with each additional
AND condition.
Likewise, the inclusion of NOT decreases the volume of output. However,
the NOT condition day also cause relevant citations to be excluded from the
output when a citation. contains the desired terminology and the unwanted terms.
Thus, both the total number of Output-and the total number of relevant citations
17
decreases with each additional NOT condition.
It is quite possible that a given search request might be viewed by two
different searchers as consisting of a different number of facets. This seems
to depend upon the language, of the topic, indexing vocabulary, and interpreta
tion of the query and indexing language by.the searcher when interviewing the
client. For example, the search topic, "AV Aids for Library Instruction of
Users," has divided into two formulations in the following example:
Formulation A Fcrmulaf ion-B
AV Aids Library Instruction Users AV Aids Instruction of Library Users
(2)
These,differences in conceptualizing the query will result in different
search formulations: The approaches to representing each concept in a-query and
the combination of conceptual groups form the central,. core of the online search
strategy process. Models of several of the more predominant approaches to
executing search strategy are shown in idealized representations in the
following discussion on approaches to search strategy. No particular se
quence IS intended, and there is no intention to recommend one approach
over the others. The object e illustrations and, discussion is to heighten
each searcher's interest in which approaches are used,and to aid in the analy
sis of which' approach yields what kind of results.
I. Building Blocks.
The first of the approaches is the building block (or subassembly, or
'modular) approach, and is characterized by an attempt to develop each aspect
of the search separately, as if it were a subsearch all on its own, and
then make the final logical assembly of all of these subprograms.
This is a predominant style:for many searchers. Individuals who are accus
tomed to break all of their problems into subproblems may use the building block-
17
approach as a natural-extension of their ow _ethod of tackling a specific
probl
BUILDING. BLOCK APPROACH
The building block approach, illustrated abbve, is demonstrated in the
following search topic on "IdentificatiOn of the Artistically Gifted." The
topic is broken down into three concept groups: GIFTED, IDENTIFICATION, and
ARTISTIC. In the table below are listed the three facets accompanied by search
.terms used to represent the facets in the search formulation.
ARTISTIC GIFTED
Art GiftedArts TalentArtistic TalentsArtistically ) TalentedEsthetic -allyEstheticsAesthetic -allyAesthetics ,
1DENTIFICATIO
IdentifyIdentifiedIdentifyingIdentificationIdentifies
Terms from each facet are then assembled into search statements joined by
the OR operation., The linal step in the building block approach is to combine
the results of the Subsearches, in this case, by the AND operator.
Formulation
1 S ART ? -?
2 S ARTISTIC?
3 S ESTHETIC?
4 S AESTHETIC?
ComMents
Search statements 1-4 form the conceptgroup ARTISTIC
Formula fon (con
5 C
6 S GIFTED
7 S TALENT?
8 C 67/OR
9 S IDENTIF?
10 C S, AND 8 AND 9
Statements 6 and 7 form facet GIFTED
Statement 9 forms the facet ,IDENTIFICATION'
The combination of the three facets by theAND operator is done in,the final-searchstatement
The building block approach has the advantage of providing a somewhat clearer
trace hi,story, one that is easy to review and understand at3a later date. It
tends to follow and read like the actual query formulation; however, advan-tagesof the approach are that more working memory and online time are used in com-
parison to the other approaches since the searcher establishes new sets with each
command and introduces the main AND conditions only at the final stage of
formulation. This approach may use more online tli e than is necessary. For
example, the union of two concept groups in a 3- acet search may result in no
or so'few citations that it would not be necessar to impose the third condition.
In that case, developing the third subprogram before nal COMBINE operation.
on the three subprograms would be an unnecessary use of tune and effort.
modification of this approach would be to perform a Brief search in order to
establish the likelihood that the combination of one tei from each of the.three
facets results in some hits.
Citation Pearl Growin
This approach makes the most use of tile interactive
search. system. The technique here is to star
capabilities of the online
h a very direct search on the
most specific terms in the search request and find at least one citation. The
searcher then calls up one o more-of these citations for online review, noting
index terms and free text found in some relevant citations. These new terms Are
19
incorporated into the ulation to retrieve additional citations. After adding
these terms to the formulation, one can again review more retrieved citations and
continue this process in successive iterations until no additional terms are found
that seem appropri.te for inclusion.
CITATION PEARL GROWING APPROACH
The Citation pearl growing approach to search strategy is illustrated,above.
The topic, 'identification of the Artistically. Gifted," is searched using-this
apprOach. The Briefsearch is first employed to retrieve a few citations for
online review.
Briefsearch
GIFTED(C)IDENTIFICATION(C)ARTISTIC
One citation is retrievedrieved and printed in format 2 so that the descriptors can be
reviewed for possible inclusion in a revised formulation.
ED10407:q EC071443THE ENTIFICATION OF ACADEMIC CREATIVE AND LEADERSHIP TALENT FROM
BIOGFHPHICAL DATA. FINAL REPORT.INSTITUTE FOP BEHAVIORAL RESEARCH iN CRE6TIVITY. SALT LAKE CITY1-
UTAH. ; NORTH CAROLINA STATE DEPT. OF PUBLIC INSTRUCTION RALEIGH. DIV.Fop EXCEPTIONAL CHILDREN.
74 76P.SFONSORING AGENCY Z. SMITH REYNOLDS FMUNDATION! SAPELO ISLAND' GA.EDPS PRICE MF$: iL 76 HC$4.43' PLUS POSTAGEDESCRIPTOPS:-ACADEMIC ACHIEVEMENT/ *BEHAVIOR PATTERNS/ *CASE. STUDIES
(EDUCATION)/ CREATIVE ABILITY' *CULTURAL FACTORS/ EXCEPTIONAL CHILDRESEARCH/ *GIFTED/ HIGH ACHIEVERS/ *IDENTIFICATION/ LEADERSHIP/ RACIALFACTORS
Substituting the descriptor phrase, "Creative Ability," for the term artistic,
the Briefsearch is revised.
Briefsearch
2 S CREATIVEMABILITY(C)IDENTIFICATION(C)GIFTED7
Five citations are retrieved; typing the retrieved citations in ormat 6
allows the searcher to browse the titles of the set in order to find candidates-
for printing more portions of the record.
1/6/1ED104095,.
THE GIFTED AND TALENTED= A HANDBOOK FOP PAPENTS WOPKING DRAFT.
1/6/2_-ED10404
THE IDENTIFICATION OF THE GIFTED
1/5/3ED104039
THE IDENTIFICATION OF ACADEMIC' CREATIVE AND LEADEPSHIP TALENT FROM
D TALENTED.
BIOGPAPHICA .BATA. FINAL PEPOPT.
1./6/4ED102773SUGGESTION- FOP IDENTIFICATION OF GIFTED AND TALENTED STUDENTS.
1/6s5ED1001 Ill
TEACHING GIFTED CHILDREN APT IN GRADES ONE THROUGH THREE.
The la'St title seems relevant CO
forMat 2.
- E1t100102 95 Fi-.07A972
he search topic so it is then printed in
TEACHING GIFTED CHILDPEN APT IN GRADES. ONE THROUGH THREE.LUCASMARK C.; ALLEN= BONNIECALIFORNIA STATE DEPT. OF EDUCATION SACRAMENTO. DIV. OF SPECIALEDUCATION.74 46P.; FOP ADDITIONAL INFORMATION SEE ED OBS 253 AND 254 AND ED.082 433'SPONSORING AGENCYi BUREAU OF ELEMENTA AND SECONDARY EDUCATION-(DHEW/OE)5 WASH1NGTON D. C:.
EDPS PRICE MF-$0.76 HC-$1.95-FLUS POSTAGEDESCRIPTORS; *APT/ 'CLASS ACTIVITIES/ CREATIVE ABILITY/ CREATIVEEXPAESSION/ +CUPPICULUM DEVELOPMENT/ CURRICULUM' GUIDES' EXCEPTIONAL_CHILD EDUCATION/ *GIFTE/ PRIMARY GRADES' *PPOGPAM PLANNINGIDENTIFIERS; ELEMENTARY SECONDARY EDUCATION -ACT TITLE V-/ ESER TITLE
r.
Additional searching vocabulary derived from this reiord may include "Creative
expression," "art" or "arts." This process could then be continued by adding new
terms to the Briefsearch formulation or incorporating search terms into a building
block or other approach.
n pe_rlcv owing is characterited by the development of the search
formulatien.in a very dynamic, empirical manner. Thinking_ time associated
this approach may result in a longer online connect time than might be achieved
with searches in which some preparatory work and Planning, is done before online
searching. Not only can online review of retrieved citations be very helpful in
the idedrification of search terms for addition or deletion, but the payoff from
this approach can be incorporated into any of the other approaches.
SUCCESSIVE FRACTIONS APPROACt
3. Successive Fractions
Successive fractions divide and conquer, or file partitioning) implies
that an initial bite of the file is made in order to 'Assemble a set that satisfies
the conditions of the first facet,of a Multi-faceted search. T is first facet may
be a portion of the-file such as year of publication, accession ange, ERIC clear-
inghouse document type or availability, major descriptor or major identifier.
When the second facet of the search query iS applied as an AND condition to the
partitioned subfile, the result is the establishment of an even smaller set or
subfile. The remaining search facets continue to be applied to the subfile resulting from the previous AND conditions,.
DIALOG's LIMITALL feature, when used at the beginning of a search, serves as
an example of the successive fractions approach. .the following example, the
LIMITALL capability is employed before the search formulation in order to restrict"
searching to the -titioned subfile, which is, in this case,..only the periodical
literature subfile.
Formulation
LIMITALL/EJ
1 S IDENTIFICATION(F)GIFTED/TI
The LIMITA
Comments-
The retrieved set will only contain the terms. "Identification" and "Gifted" in the title
of a journal article.
feature can alsotte used to restrict doenMent year by using the
accession iber range. Since ERIC ONTAP contains only 1975 ERIC accessions, the
LIMITALL capability for document year has already been Imposed.
The successive fractions approach is also employed to restrict volume
output. A good example of this technique is found in search topic 113, Evaluation
of Primary School Reading Materials. Since the combination of all three concept
groups, REAPING, EVALUATION, and PRIMARY EDUCATION retrieves over 300 citations,'
the volume of output is trictedeby limiting search terms to occurrences inthe
tie, identifier, and descriptor fields. Output can also be limited by clearing
house, document availability or type,
Using,the successive fractibns approach, the searcher can work with sets
that become progressively smaller, and the search can be terminated whenever the
size theeet is satisfactory to the searcher. For example, in the following
search formulation, 45 citations are retrieved after the application of the second
facet. Instead of imposing the third facet, ARTISTIC, and risking the loss of
some relevant items, the searcher can terminate the search at this point or limit
the output, as in this case, by clearinghouse.
Formulation Comments
1
2
4
5
6
S IDENTIF?
SIALENT?? ?
S GIFTED
C 1 AND (2 OR 3)
S-CHEC
C 4' AND 5
45 Citations are retrieved at this point
Rather than impose the third,facet tothe set, the searcher restricts theoutput to Exceptional Children Clear-inghouse. The output contains 26 ci-tapionsat the conclusion
The major advantage with this approach is that if permits the search to be
completed at an e slier point than would be the case with the building block
approach if the size of the retrieved set fulfills the search objective before
exhausting the search facets. The successive fractions approach also eliminates'
the necessity of "backtracking" or reformulating the search. strategy when the
final set fails to satisfy the search objectives.
The Boolean NOT operator is used in Che formulation below to demonstrate how
the successive fractions approach'is employed to create a facet AMERICAN INDIANS;
since the Indians of India are not wanted in the retrieved set, it is easier to use
the NOT operator rather than keying-in separately all American Indian tribal names.
Formulation
1 S INDIAN
2 S INDIANS
3 C 1 OR 2
4 S INDIA
5 C 3 NOT 4
Comments
Statements 1 to 3 create a set containingall occurrences of the terms Indian orIndians. Truncation is not used be-cause of the likelihood of retrievingfalse drops like Indiana or Indianapolis.
Statement 4 creates a set containing alloccurrences of the term India. It isexpected that documents. containing thisterm would treat the subject'ef theIndians of India, a topic not desired.
The set of documents in statement 4 is"Subtracted" or'partitioned from theset created by stateMent 3.
The NOT operator serves as a good illustration of how the f le is virtually
partitioned and attention is paid only to the newly assembled portion
successive fractio approach. This approach is often used advantageously with
such as MOSIS CA Condensates and Agricola as thes 1 s have major sect
headings that are used in tha,printed indexes, which when used as search terms at
the beginning of a search, form a subfile for detailed subject searching.
MOST SPECIFICFACET
,,optional
MOST SPECIFIC FACET FIRST APPROACH
4- Most Specific Facet First
This.fouith apprOach consists of starting a multi-faceted search formulation
with the mostmspecifle aspect of the query that is not likely to suffer from an
vagueness of indexing (e.g., a named object such as'Foreign Language Test, or
Boy.Scouts If the postings are sufficiently low after scanning the results,
the searcher may choose not to imposethe remaining Conditions-:to thaaeareh
unless higfi precision performance is a search objective:
In search topic #2, "4-H Clubs, their members and activities,"
the search formulation is comprised of variant spellings' of the club'. Since
21 citations are retrieved using' free text and 14 citationssearching_ -e
fier field, the search is terminated without applying the members"
aspect of the search as the final set, in both cases, is sufficiently
If there had been too many citations after searching the most specific facet,
then the next most specific facet ould be searched and combined in an AND rela-
tionship with the results obtained with the first facet. The resulting inter-
mediate result may now be small enough that there is no need to impose the rest
of the search conditions. This approach can-be more efficient than the full
building block model when a search topic has some very specific facets with low
postings.
LOWEST POSTINGSFACET
OptiOnal
optional -
LOWEt '0 TUGS FACET FIRST APPROACH
-west PostipaE_Eacet First
The fifth approach starts a multi- faceted search with the facet that is
estimated (e.g., fro frequency search aids, 7th edition ERIC Thesaurus,
or from experience) to have the smallest number of postings. If the postings are
fficiently low there may be no reason to continue searching additional facets.
For example, a searcher may safely assume that the facet GIRL SCOUTS OF' AMERICA
derived from a topic on "Success rates in Higher Education of students who were
members of the Girl Scouts," has fewer postings that the facet HIGHER EDUCATION.
-Since the search for the facet GIRL. SCOUTS results in less than 20 citations, the
rcher terminates the search before applying th second facet HIGHER EDUCATION
to th formulation,
wary
There are other approaches to interactive searching, and of course, it is
always possible to work with combinations of the above approaches. There is
little in the,-ay of solid theory or guidelines regarding how a search should be
formulated.
27
STEP e4
COMPILING THE SEARCH TERMS
Before going online, some check of word lists and thesauri will help in
that the search terms used will lead to useful results. Both the controlled
ure
vocabulary a data base (indexing recards indescriptor/identifier fields)and
the free text (subject-conveying fields other than descriptor and identifier fields)
are possible sources of search terms.
Search Terms in Descri.tor and Identifier Fields
In the ERIC data base, the descriptor and identifier ids contain te n-
ology which has been assigned by ERIC Clearinghouse staff. The Thesaurus of ERIC
Descriptors _ an example of a controlled vocabulary which is used by the indexer
in ERIC clearinghouses. If the Thesaurus is also used by the searcher the in-
tellectual effort required to compile search terms can be reduced as it serves
as a control on variant word forms, synonyms and homographs. For example, the
ERIC Thesaurus suggests a number of search tetras to represent the concept of,
'violence: crime, aggression, antisocial behavior, delinquency, emotional response,
and the descriptor, violence.
The identifier field contains names of persons, places, things, which are
important terms associated with the Item. No controlled list like the Thesaurus
used by the indexer when identifiers are noted, but there are some rules ,for
entry which control this operation, and each type of identifier is now being
reviewed and'ehecklists compiled by the ERIC Network.
2. Free Text Searching
The free text _c ing capability of DIALOG allows for the scanning of all
subject-conveying fields of the ERIC record: title, abstract, descriptor, iden-
tifier, descriptive note, corporate source'and sponsoring agency. If a concept like
29
parapsychology is not found in the ERIC Thesaurus, terms synonymous with
parapsychology can be gathered from such lexicographic sources as the Thesaurus
of Psychological Terms, Roget's Thesauru , an unabridged dictionary and other '44
sources. This list of search terms compiled from sources other than the ERIC
Thesaurus can be searched as free text in all, or a specified number of subject-
conveying fields; however, recall scores may be low if only a single subject-
-conveying field is searched.
As citations are retrieved using this free text scan method a few citations
can be called up for online review to note how relevant citations are indexed.
These descriptors and other ;free text terms can be used in further search formu-
lations'= This meth 'od of gathering additional search terms resembles the "citation
pearl growing" approach to the formulation of search logic discussed in the p
vious step of this section. Titles of known relevant items are also a rich source
of "free tex search terms.
3. Online Alphabetic Word List and Thesaurus
When a planned search has more than one term associated with it that occurs
close to ether alphabetically, it is fast and convenient to review the online
dictionary (an alphabetic list of words in all the searchable fields of the ERIC
data base). Rather than directly keying-in the search terms in a series of
SELECT commands the EXPAND command is used. Additional terms may appear which
had not been considered previously.
Entries like corporate sources and author names are particularly appropriate
author and corporate source names and because typing errors may occur when keying-
in long corporate source names
corporate source names.
Scanning the online thesaurus, the subject-related descriptor list, may
suggest to the searcher during the online activity search terms which do not
,appear in alphabetic proximity, The postings for each term are given and this
30
data can afLect the search strategy being considered. Heavily posted terms may
be selected after reviewing the EXPAND display of a lightly posted term, for
instance, or vice versa. These strategies are discussed in the previous step=
However, selecting a phrase from the online thesaurus or alphabetic word list
restricts retrieval to occurrences of the phrase as a multi-term descriptor
or identifier. Retrieving single and multiple word search terms from all
subject-conveying fields is covered in more detail in step 6.
STEP 5nimism~
ORDERING OUTPUT
Ordering output occurs not as a single, isolated event in the sequence of
online Search activities. Rather, because decisions influenced by output spe-
cifications are made throughout the search process, the actual activity of
ordering output is not performed t any fixed moment Curing the online search.
As soon as the Searcher begins formulating search strategy, the construction
is affected by conditions imposed by the requestor concerning output spe-
cifications, e.g., size of output, document type, year of publication, etc.
For example, in order to fulfill the request of an information seeker who
desires a all'amount of output with a high_ percentage of relevant items, the
searcher may select the "citation pearl growing" approach to search strategy.
The LINITALL command may be preferred by a searcher who must restrict the re-
sults byspublicat n date. In another instance, the Briefsearch and a shortened
online print format may be favored When the information seeker is hamperdd by
time restrictions and requires immediate results; a more comprehensive search
may be performed at a later time.
Just as ordering output influences the planning ofsearch strategy (step
3 in t e process of online searching), it affects the conceptualization of theLsearch as input to the ret -efal system (step 4). If the information seeker is
concerned about retrieving all relevant citations at a minimum cost, the
searcher is compelled to elect an approach which will result in high
and formulate the strategy employing time-saving devices, e.g. EXPAND and
range SELECT, low activity search time, pre-search-offline preparation to
minimize online-"think" time, etc.
Output can be scanned online fine tuning or for further trimming of the
search results on the basis of non-subject parameters. For example, the resul-
tent setjmay exceed the output size requested by the information seeker so that
the searcher, depending upon policy, personal judgment or consultation with the
requester may decrease the size by applying limiting criteria such as language,
date or type of publication, availability, sponsoring - agency, etc. These spe-
cifications are in most cases applied to an existing set. The decision to trim
the output may be the outcome of the Preliminary evaluation (step 7), in which
the search res =ults are reviewa-d te= determine whether the search strategy sat-
isfies the search objective.
The retrieval system and/or data base plays a role in- determining 7711aracter-
istics of the output. Sorting criteria vary among data bases; format of output
not only differs from DIALOG data ba,e to data base, but DIALOG offers pre-
determined output formats rather than allowing the searcher to structure his/her
own format.
In- summary, search output can be limited by criteria imposed by the infor-
mation seeker, searcher, or retrieval system. The influence of ordering output,
prevails throughout the search process; however, ordering output, as a step-in
the process of online searching, exists as a concrete and very real activity,
but May occur at a-number of points during online'inte action.
34
STEP 6Immt
CONCEPTUALIZING THE SEARCHAS INPUT TO
THE RETRIEVAL SYSTEM-
The features of any computerized retrieval system influence the way in
which a given search is formulated for that system. Only a few of the operations
which can seriously affect search strategy operations are described briefly below.
For a more complete description, see A Brief Guide to DIALOG Sea hin-
C. Bourne, DIALOG Lab Workbook.
1. Word Proximit Searchin for Sin le Term Combinations
The word proximity feature can be used as a shortcut to exp ess a direct
logical AND combination. For example, a Briefsearch employing word proximity
would achieve the same search results as the more conventional, sequential,
formulations in the following search topic on the effects of TV violence on
Se u n lel Search formulation
1 S CHILDREN
2 S TELEVISION
3 S VIOLENCE
4 C I-3/AND
ord Proxim
children;
Search Formulation
1 S CHILDREN(C)TELEVISION(C)VIOLENCE
There are some disadvantages with the word proximity approach. It does not
provide any intermediate sets. In the event that there is an interest in modifying
the search formulation. it cannot be used with truncated words.*
The use of the word proximity feature with its implicit AND operator is
the preferred practice for searches, or parts of searches, when:
* this will be-changed in DIALOG very soon.
the resulting command statement is a single typed line in totallength, and
2) the search terms are lightly posted.
Word proximity is partic.slarly advantageous when the search is planned to be-
made over all rather than a few specified fields.
Mu _e-Word Search Ter - Word Proximity within a Fie
the Record
In the case of single word search terms, DIALOG scans all subject-conveying
fields by default unless otherwise directed by the searcher. Multiple-word terms
entered by the searcher are only checked against the entries from the d±§:cIlE9s=
or identifier fields. Searchers restricting their searches to the controlled
vocabulary fields of the ERIC record (descriptors and identifiers) are ignoring
other subjec -conveying items of the record. High recall is achieved when the
searcher uses both the controlled vocabulary and free text. Searching can be
extended to all subject-conveying fields simply by using the adjacency feature (W)
as shown below.
Search Formulation Result Effect
S BOOK CATALOGS85 Searches in descriptor and identifier
fields for subject headings thatinclude the character sequence,BOOK,CATALOGS
95 Searches as above, plus title, abstract,descriptive note, corporate source,and sponsoring agency fields for theoccurrence of this character string
S BOOK CATALOGS
The use of the wcrd proximity feature in this way generally retrieves about
20% more postings than are retrieved by searching fields with assigned index terms.
The significant increase in postings results in a high recall search. But remember,
more computer time is spent,using word proximity, especially for high posting
combinations like ELEMENTARY(W)EDUCATION.
36
There are many instances in which the controlled vocabulary does not yet
include current expressions that may appear in the literature, The word prox-
imity feature can help find chse.citations by means of direct text searching
as in the following examples:
A. Search by Controlled. Terms Results B. Ward Proximity Free Text Search
A B
S TAXPAYER REVOLT 0 S T'AYER(W)REVCLT
S INFORMATION EXPLOSION 3 '40 S INFORMATION(W)EXPLOSION
S STREET PEOPLE 1 4 S STREET(W)PEOPLE
S BACK TO BASICS 12 59 S BACK(1W)BASICS
S GENERATION OF ALTERNATIVES 0 S GENERATION(1W)ALTERNATIVES
3. StaraciteraarLintitatitmgThe search formulation can be influenced by the amount of working storage
available to the searcher. About one million postings per search are availablery
to each searcher. Computer memory limitations pose less of a problem in the ERIC
file than in large files such as BIOSIS or AGRICOLA. Since problems do occur,
certain formuintion strategies can be used to reduce the amount of storage space
needed= Storage-saving techniques that can be used are:'
a) Avoidance intermediate logical formulations
b) Stem search
c) I '.,:PAND and range SELECTS
d) Word proximity feature
WIT feature:
f) BEGIN, resetting set numbers to zero
g) Search Save
a) Intermediate results of COMBINE operations that are stored online
take up space in the same way as the other gets. Combining all the logic in. .
.one statement reduces storage space. Consider the following example in which_ statement_
37
the same search objective is met two different ways, but -h differences in
the total amount of working storage space required:
SeArch Plan: (TEST? or EVALUATE? or ANALY?) and (READING or MATli _TICS)
A Formulationo- ulation
1 38050 S TEST? 1 38050 S TEST?2 49153 S EVALUAT? 2 49153 S EVALUAT?3 47022 S ANAL?? 3 47022 S ANAL??4 100632 C 1-3/OR 4 21461 S READING5 21461 S READING 5 10520 S MATHEMATICS6 10520 S MATHEMATICS 6 14598 C (10R20R3) AND (40R5)7 305078 14598
C
C
5-6/OR4 AND 7 -(180,804 postings in working storage)
(311,943 postingsstorage)
in working
Performing the same search two different ways requires a-difference of 131,139
postings in working memory, which could be important for some searches. The pre-,
(erred practice is to avoid the intermediate logical operations if a storage
overflow problem is anticipated.
h) If a ward is sufficiently unique, the use of a word stem instead of the
combinetien of a series of terms might save a small amount of storage space, be-
cause an intermediate logical operation might be avoided through the implidit OR
operation performed in the stem search. Consider the following'example in which
the same search ob ective s'met two different ways:
Formulation A Formulation B
1
2
3
4
5
6
7
3874 S
3680 S
4024 S
42630 S
2373 S
2016 S
48737 C
EVALUATEEVALUATEDEVALUATINGEVALUATIONEVALUATIONSEVALUATIVE1-6/OR
postings in working storage)
1 49153 S EVALUAT?postings in working storage)(49,153
(107,334
In formulation B, over 58,000 storage spaces have been saved by using
stem searching. The preferred practice is to use ;he word stem
suffiCiently unique not to cause too many false drop problem
when the word is
c) The EXPAND and range SELECT feature can be very helpful with storage
problems because of the implicit OR operation that is done with the range SELECT
feature. This can be done for both the alphabetically-related terms and the
thesaurus - related terms. In the following example, the same search objective is
accomp 'd in two different ways:
(AUDIOVISUAL EQUIPMENT, FACILITIES) and (LIBRARIES of all types)
Formula n Bion A
1
2
3
4
5
6
7
8
6,889-4,796
4,3432,694
1,5561,102996995
S TELEVISIONS FILMS
S AUDIOVISUAL AIDSS INSTRUCTIONAL MEDIAS INSTRUCTIONAL AIDSS AUDIOVISUAL INSTRUCTIONS SLIDESS TAPE RECORDINGS
E(AUDIOVISUAL AIDS)20,016 SR1-R18,R2O-R31,R35-R37,R41
E(LIBRARIES)2 9,241 SR1-R13,R15,R22-R243 1,469 C 1 AND 2
(30,726 postings in working storage)
9 800 S INSTRUCTIONAL FILMS10 469 S EXHIBITS11 426 S VISUAL AIDS12 338 S PHONOTAPE RECORDINGS13 265 S PROJECTION EQUIPMENT14 2-49 S CARTOONS15 226 S PHONOGRAPH RECORDS16 191 S TAPE RECORDINGS17 152 S VIDEO CASSETTE SYSTEMS13 79 S SCREENS19 75 S REALIA'20 52 S DISPLAY PANELS21 40 S MICROPHONES22 19,040 C 1-21/OR23 6,802 S LIBRARIES24 2,310 S LIBRARY SERVICES25 1,129 S PUBLIC LIBRARIES26 1,173 S UNIVERSITY LIBRARIES27 1,045 S INSTRUCTIONAL MATERIALS CENTERS28 801 S SCHOOL-LIBRARIES29 740 ,S INFORMATION CENTERS30 529 S SPECIAL LIBRARIES31 371 S ARCHIVES32 289 S STATE LIBRARIES;33 230 S RESEARCH LIBRARIES34 197 S NATIONAL LIBRARIES35 7,985 C 23-34/OR36 1 048 C 22 AND 35
(69,478 postings in working storage)
In thin
working mi
mple the EXPAND/SELECT approach saved almost 40,000 postings in
lo eover, formulation A took more time to type, increasing online
connect tiow And increasing the possibility of typing errors.
d) The word proximity atu re, discussed earlier in step 6, does not provide
any interrediary sets. 1f the searcher intends to revise the formulation during
the online interdct ln, using word proximity as a means to bypass computer memory
ay be suitable to expresslimitations is not suggested. However, word proxi
a Brie there is no need for intermediate sets. No-e in the following
example on the eft of TV violence on children that over 47,000 poStings are
saved using word proximity.
Formulation A
1 38.189 S CHILDREN
2 7,930 S TELEVISION
3 984 S VI6LENCE
4 133 C 1-3/AND
(47,23.6 postings in working memory)
Formulation B
33 S CHILDREN C/TELEVISION(C)VTOLENCE
postings in working memory)
e) rse the BEGIN command when doing several unrelated searches at the same
terminal session. This resets set numbers to zero, and clears all of your working
storage space, thus starting subsequent searches with a full allocation of working
storage.
1) The LIMIT feature can be used to do some tasks thWMight otherwise4
require more storage when done with set combinations. This is particularly true
when restricting the search output to accession year or document type, as shown
in the following example:
Search Plan: (VOCATIONAL TRAINING) and (ERIC accession YEAR -4 1975 or more recent)
Formulation LIMIT CommandFormulation A
11,615 S VOCATIONAL(W)EDUCATION 1 11,615 S VOCATIONAL(W)EDUCATION
2 32,120 S 11175 2 3,768 LIMIT1/095254-999999/ED
3 37,145 S YR=76 3 1,015 LIMITI/101873- 999999/EJ
4 36,369 S 4 4,783 C 2-3/OR
5 12,317 S. YR =78 (21,181 postings in working storage)
6 4,706- C 1 AND (2 OR 3 OR 4 OR 5)
.(141,272 po tings in working storage)
In this case, the most convenient and fastest technique for the searcher is
to use the YR= feature instead of looking up and keying'-in the ED- and ET-
number ranges for reports. In the above case over 120,000 storage positions have
been saved.
hen storage problems are anticipated, the LIMITALL feature can be used
to partition the file by accession number or document availability. The search
is performed on the portin'of the file specified in the LIMITALL command.
Using the same search plan above, the LIMITALL capability is demonstrated below:
Search Plan: (VOCATIONAL EDUCATION) and (ERIC accession YEAR = 1975 or'=more recent)
elation C; LIMITALL Commands Comment
LIMITALL /095254- 9999999/ED
1. 3768 S VOCATIONAL(W)EDUCATION
LIMITALL/101873-999999/EJ
2 1015 S VOCATIONAL(W)EDUCATION
3 4706 C 1-2/OR
LIMITALL/ALL
(9412 postings in working memory)
The LIMITALL command restricts the file to1975+ ERIC non-journal accessions.
Occurrences of the descriptor phrase areretrieved from this partitioned ,file
The new LIMITALL command cancels thefirst command.
The LIMITALL command is removed in this wayto permit searching in the entire file.
Although the LIMITALL feature keeps the number of postings in the working
- memory at a _imum'when comparing the results of the three strategies above, one
disadvantage of the LIMITALL feature is that the searcher is required to repeat
the search statements (in the above example, one phrase, VOCATIONAL(W)EDUCATION,
is entered twice) if both the journal and non'-j urnal citations are desired;
however, LIMITALL is a powerful storage - saving device if one of the output spec-
ifications is year of publication or ERIC document type:
g) The search save is the most powerful storage-saving technique of all.
A search save can be prepared and saved for frequently occurring facets and exe-
cuted when needed. An example of its use is shown in the sample formulation.below:
Search Plan: (READING or READER) and
Search Save Formulation
i.EMENTARY.or SECONDARY EDUCATION),
1 22793 S ELEMENTARY(1W)EDUCATION
2 30804 S SECONDARY(W)EDUCATION
3 25949 S READING?
4 4458 S READER?
5 . 6898 C (3 OR 4) AND (1 OR 2)
END/SAVE
ommen
The Search Save command (END/SAVE) con-cludes the routine. DIALOG assigns a-
SERIAL#1EAA serial number to the Search Savo and-23MAY78 8:52:56--DSER 4111 the searcher can resume searching.
$0.00' 0.046 HRS FILE1 00` DESCRIPTORS$0.37 TYMNET$0.37 ESTIMATED TOTAL COST
Search Plan: (READING or REA6ER
Formulation.
.EXECUTE-1PAA
2279330804-
2594944586898
68Q8-
and (ELEMENTARY or SECONDARY EDUCATION) and COSTS
Comments
ELEMENTARY(1W)EDUCATIONSECONDARY(W)EDUCATIONREADING?READER?(3 OR 4) AND (1 OR 2)
SERIAL// 1PAA
6577 S COSTS
91 C LAND 2
In the example above,
the first formulation are erased and the searcher uses only 6898 storage locations
The Search Save is recalled and performedby the system at the signal-of the.EXECUTE command
The system responds to the .EXECUTE commandby printing the results of all thesearch-statements but only -the finalstatement retains the record of post-ings. In this, case, it becomes state-ment .1 of the search in-progress.
The third facet is applied to the fo ula-tiun
The search.fresults
the 90,902 storage locatione'nsed in statements 1-6'of
for the two facets READING and EDUCATION.
Such a search tactic can also be temporarily constructed and used during a
terminal session. The Search Save has the effect of eliminating all tem-
storage:space for the development of the two facets READING and EDUCATION,
-then be released as soon as the rest of the search has been completed.
It should be noted that the Search Save approach can help with the storage
problem, but xtra computer time is required to reprocess all of the search steps
41
which comprise the saved search':
The appendix to this report includes some Search SaVes for common educational
level facets. The procedure for recalling and exchanging these Search Saves
between workspaces is covered in the introduction to the appendix..
4.- Searchin= Burin= 0 Activit Times
To reduce the cost of searching, the search should be performed during the
system's low activity hours. Most online seara systems have a respon e time,
and hence a total search time, that depends upon the volume of processing activity
that is taking place at that time. The greater the number of simultaneous search
the slower the response time. Searches done during the 8-2 (PST) time slot are
generally slower, since all U.S. time zones are in the course of their normal
king day. Search speeds for the same search can sometimes differ by a factor
is made. A search that takesof 2, depending upon the time of day that the search
five minutes to run at 3 P.M. (PST), 'during the low activity period, could some-
times take 10 minutes to run at Noon (PST), a high a t vitYperiod.
STEP :-7sm.mmailmo
EVALUATINGPRELIMINARY
RESULTS
undamental to the evaluation of the search strategy is the question,
"How well is the search formulation meeting the information need?" if' -it is
determined in the original interview that the information seeker wants highrecall, then an appropriate performance measure would be a check on the fractionof relevant citations in the file that are actually retrieved by the search.if the objective ts only a few relevant items at a low cost to the information
seeker; then search time would be an important measure and a Briefsearch maybf satisfactory if some relevant items are retrieved. While online-, the searcherShould monitor the search in progress and check on results. This check mayrequire a change in strategy and overall approach to the search.. A review ofSteps 1-6 may be necessary.
STEPnomMirMINIO
EVALUATING FINAL RESULTS
There is no sin le measure of ho ood a search is. Performance measures
such as cost, recall, and precision can be applied to the search results given
the specific search objectives:
Search Ob ective
* To retrieve all (6r as many aspossible) relevant citation' infile
* To reduce (or keep low) the numberof non-relevant citations, retrieved
* To-keep costs low and to retrievesome relevant citations
1. Recall
Recall is
hard to find
Certain tasks can be perfo -med in
done on both ._assigned index fields and
erformanc Measure
Recall (defined as percent of relevantcitations in the file.that are actuallyretrieved)
Precis__n (defined as percent of retrievedcitations that are relevant)
Cost (defined as connect time, and cost,of offline'printing)
easy to define, but very difficult to measure because it is so
the total number of relevant citations that are in the file.
order to- achieve high recall-. A search
titles will generally return more cita-
Lions (total and relevant) than a search done on one of those'fielda.themselves.
The larger the total number of subject fields of the bibliographic records being
searche::, the larger the total number of returned citations.
Precision, sometimes called relevance, is easy- to UnderStand and easy to
measure. Whereas recall increases as more parts of,a citation a arched,
*The search questions provided in the next section have all been searched severaltimes by several different people, and the output reviewed, so that there is afairly good idea of what citations in the file are relevant to each of the ques-tions. This information has been stored away as answer sets, so that it can be-retrieved to permit the practice searcher to compute recall.
conversely, precision decreases as more parts of a citation are searched. If
a low volume, high precision
scrip
search is desired, searching the title, major de-
rs. and major identifiers delivers a high percentage-of relevant items.
3. Cost
Connect time or computer cost is also easy to understand, and easy to
measure because the accounting information is provided as part of each search
output by the use of the BEGIN, END, .COST or LOGOFF commands. Using the time-
saving schemes which have been described in Step 6 can serve to aid the searcher
who desires to lower costs of online searching.
SECTION II
ERIC ONTAP EXERCISES FORSELF-IMPROVEMENT OF
ONLINE SEARCHING
INTRODUCTIO TO DIALOG'S ERIC 0 TAW FILE 2013
RIC ONTAP is 'a collection of RIE and CUE citations from the 1975 ERIC
file. Its contents correspondto citations In all 1975 printed ERIC indexes, n
necessarily to a 1975pnblication date,.and include the range of am secession
numbers EJ 101 873 to EJ Iv 926, and ED 095 254 to ED 110 594. ERIC ONTAP is
exactly the same in all respects (data-elements, searchable fields, ttc.). as the
1975 accessions in the regular DIALOG ERIC file (file-4).. -
ERIC ONTAP represents.12% of the total ERIC data base (some 35,400 items
There are points to remember when using the file:
Since the file is small, you will not encounter problemsinvolving size, 'e.g., what to do when you still have alarge number of posting n your final set.
Spven editions of-the.to index file 1, but oERIC ONTAP indexing.of terms over the .periodin ERIC -ONTAP to the sam
RIC Thesaurus,were used in ordery the fifth edition was used fors, shifts in meaning and usage
several years do not occurextent as in file
The, unusual feature of the ERIC ONTAP file is that it contains "answer sets"
(lists of ERIC accession numbers) to, predetermined queries. Searchers using this
file can perform a search for the query and compare-their results with the "answer
set" for that query.
The search.querie6 represent
asSembled from questions posed to
real,,not synthetic questions, since they are
erence librarians and :ERIC Clearinghouse
searchers. Level of difficulty varies from query toquery and they are graded
simple, moderate, and difficult. "Named object" queries, e.g., 4-H Clubs,
16 Personality Factor Test, are easiest to perform but they involve problems.
fdr the online searcher which, if worked out, will help. in more complex search
formulatiens.=
This section of this manual includes 10 of the 29 queries in the
ERIC ONTAP,file. We have chosen queries of special interest, to the 16 ERIC
Clearinghouses. It is suggested that Clearinghouse searchers review-some or
all of SECTION II search topics rather than the single topic of interest to
their clea inghouse in order to execute online search operations described in
SECTION I. For demonstrations or workshops, searchers may want to use specific
q ule ries to show how online searching of the ERIC data base can be performed and
to\point,.out how differences in search strategy retrieve different output.
Each query in this section will include various search formulations h\
explanations. The way in which the searches are executed will vary somewhat
from: what would likely be seen'in actual practice. This is done solely for
clarity'of presentation. It is assumed that the searcher will be working in an
interactive manner, using EXPANDS, reviewing sample citations, and restructuring
the-strategy as forking results are obtained. Sample earch formulations are
given here to show the critical elements of the final formulation and to,empha-
size how a change in search obIrctive (i.e., high recall, high precision, Brief-
search), can effect retrieval results.
In many cases, the search formulations given for each query are composites
that were assembled with hindsight after viewing the results of several formula-
tions. (Executing the "perfect search" at one sitting is still elusive.)
The "answer set" is a list of ERIC documents that are judged to be relevant
to the topic, which are cited by ERIC accession n_ber. The set can be selected
online by using the ens er set number, g., S AN=D05. Relevance judgments for
each query are made on the basis of the, total record in ERIC ONTAP, not on the
basis of. examination analysis of the full document.or article, and are open to
qu'estion and modification. Broad review publications, e.g;4 bibliographies of
47
dissertations in educ ation,. are not judged as relevant unless there is specific
mention of the search topic in the abstract.
Since the 1 Crelevant citations, the "answer set is compiled from
the results of several, searches o each topic, it represents an approximati r
to the total number relevant citations contained in the entire ERIC ONTAP
-file, (There may be additional rd levant citations in the file for most of these
-questions. We would appreciate hearing about them from those of yod who find
them.) The 'answer sets". to search topics 11-15 are not contained in the ERIC
ONTAP file but given at the conclusion of each discussion.
For each qUery in this section, one search formulation, specifically de-
signed as a high recall search, retrieves all the relevant cirationG listed
the conclusion of the discussion. In addition, one search formulation, in
most cases, is singled out as being preferred in the light of the information
need and search objective stated in the query description.
You, the reader /searcher, should try your own search strategies and compare
your, output and your choice of-earch vocabulary with that of the high recall
search.
The Briefsearch is also given for almost every query along with an example
. of a high precision-formulation.
All the approaches to search formulation, e.g., building block , most
specific aspect first, etc., are covered in one or more queries. For
.example, the stated search objective accompanying the query on parapsychology
requires high recall. Consequently; the formulation which achieves high recall
is. selected as the preferred routine. 'Do remember, however, that no single search
strategy or approach to search formulation is deemed the best in all cases.
To- obtain an explanation of the features of the ERIC ONIAP'file, execute a
BEGIN 201 command. Typing ?ONTAF produces the following directions:
51
ONTAP OTILINE TRAINING :1, RRACTICE) FILEPROVIDES TEST QUESTIONS I:I, RELEVANTCITATIONS FOP EACH- QUESTION. THE 1975ERIC FILE IS THE SOURCE OF THE ANSWER:ETS. QUESTIONS .AND-ANSOEF SETS ARELABELLED A TO cnmpLExITY,E.G. Ah=S0:49Pri1104,AN=D0aTENOTE THE ANSWER SETSFOP THE 1FTI SIMPLE, 4TH MEDIUM AND 3RD-DIFFICULT PLIE TIONg AFTER FINISHING THESEARCH, THE PEPFECT ANSWER SET .FOR THATQUESTION CAN BE SELECTED tY crimmAtos OFTHE FORM 'SELECT AND SELECTEDANSWER SETS CAN BE COMPINED IN AN 'AND'RELATIONSHIP WITH YOUR FINAL SEARCH SETTD FIND OUT HOW MANY RELEVANT CITATIONSYOU FOUND. A COMMAND FEVAL COMPUTES THERECALL ANTI PRECISION OF YOUR SEARCH.SEE 7TORSAM FOR SAMPLE SIMPLE SEARCH
?-1APIM FOR LIST OF SIMPLE QUEST.?TARMED FOR LIST OF MEDIUM QUESTS.TAPDIF FOP LIST OF DIFFICULTQUESTIONS-EVFIL FOR EXPLANATION OF PRECFPEL.
There are additional instructions which explain features of ONTAP in
greater detail. Instead of listing these texts below, they are given as needed
in discussions ,of the search queries which follow.
It is suggested -that you start by trying the first query on parapsychology.
Read the text, develop your t ategy, and perform an online search. Then
-figur- the recall and- precision ratios of your performance; this can be done
by comparing your output with the list of ERIC accession numbers given in this
-workbook, or, selecting. the answer set for the particular query and using the
IEVAL--command available in the ERIC ONTAP file. The /EVAL command will autom
ticany compute your recall and precision ratios.
The discussions of the parapsychology query and the fifteen other.topies
_are arranged according to steps 3 to 8 of the online searching .process outlined
in Section I. The steps are summarized below:
52
Step 3. Formulating basic search logic -- planning search strategies
Step 4. Compiling the earch terms
Step 5. Ordering output
Step 6. Conceptualizing the search as input to the retrieval system
*Ster,7. Evaluating preliminary results
Step S. -Evalua ng final results
*Since step 7 affects search formulation and occurs during the onlineinteraction, the discussion 6f this step is treated, fur the mostpart, in the Comments section of search formulations.
We recommend you.begin the ONTAP exercises with the first query on para-
psychology. The discussion of this query focuses on both the steps of the online
Searching process and the procedures for using ERIC ONTAP. The discussion of
O'arapsychology will familiarize you with the features of the ONTAP file so that
you can prepare your own search strategy. The remaining queries follow the same
pattern of discussion, butthe instructions for online features of the ONTAP
file are-not covered in as much detail in subsequent discussions.
SAMPLE SEARCH TOPIC
PARAPSYCHOLOGY
Top ,Parapsychology. (S ANS09)
-Search Objective: All information in ONTAP on.this topic'
Step 3. Pornulatin3 Basic Search Plannin: Search_ trace
The search topic, parapsychology, is depicted by a single concept or face
the term parapsychology, its variant forms and synonymous terms
psychic, extrasensory, are incorporated into the search strategy. All termi-
e.g., ESP.
nology chosen to represent the facet are joined togGther by the use of the OR
operator, thus forming a single set to describe the concept.
Step 4. Comilin= the Search Terms
Usually the search terms will come from the ERIC Thesaurus; however in
this case, parapsychology is not used as an ERIC descriptor nor are related terms
such as ESP, psychic, extrasensory, etc. Consequently, other aids for collecting
searching vocabulary may be consulted such as Rot's Thesaurus or Thesaurus of
Eachologlsal Terms. Since no ERIC descriptors describe the concept of par psy-
chology, searching only the descriptor. field.of the ERIC record would be useless.
The identifier field of ERIC resumes .may contain these terms, as it`is comprised
of terminology which has been assigned to the document by ERIC clearinghouse
Staff in order to indicate the document's 'intellectual content.
step 5. Orderirt-uto
There is no mention concerning limiting the output by publication type,
--availability, etc. When selecting an approach to search strategy, the searcher...-
must take into consideration the search objective, to retrieve all information on
this topic. In this case, a high recall formulation is preferred; moreover, the
searcher should make every effort to compile a comprehensive list of search terms
in order to meet the search objective.
55
'Step 6: c2EziplEgizin- the Search as to the Retrieval SyELn
Search strategy is illustrated in the following diagram. The tee egle
contains all records stored in the ERIC ONTAP data base. A shaded circle is
used to represent the set of documents in which the concept parapsychology
occurs.
ERIC ONTAP Data Base
Parapsychology
In the next diagram, the/shaded circles indicate not only the sets of
documents formed by search t 'erms, but the retrieved portion of the data base.
The illustration depicts formulation #1 in which the controlled vocabulary'
fields are scanned for occurrence of the search terms.
ERIC ONTAP Data Base
earthing the assigned index term fields is limited to the identifier field
since tbety are no ERIC descriptors to represent the concept. Since identifier
words/and phrases describe the intellectual content of the document, this search
is expected to retrieve a small number of non-relevant items, thus it is a high
precision formulation.
Formulation #1 CommComments
PARAPSYCHOLOGY/ID
2 S ESP /ID
3 S
4
5
6 C
EXTRA SENSORY PERCEPTION/ID
EXTRASENSORY/ID.
PSYCHIC/ID
1-5/OR
Two different spellings ofsensory are entered:in order to-retrieve Vnriant forms
Abbreviated Orm of COMBINE lsOR 2OR 3
Since the search objective is to retrieve all information on,this topic,
Formulation #2 is constricted to search all other fields containing subject
describing words nd phrases. This, formulation employs the same terms as the
previous search but is-not restricted to searching only the assigned index ter
fields. The strategy utilizes the "free text searching" of the DIALOG retrieval
system in which all fields containing subject terms are scanned: title, abstract,
identifier, corporate source, sponsoring agency, descriptor, and descripttve note.
In addition, word proximity and truncation features are incorporated into the
strategy in order to retrieve word variants.
on ti2
S PARAPSYCH?
2 S ESP
3 S EXTRA(W)SENSORY
4 S EXTRASENSORY
5 S PSYCHIC?
6 C 1-5/OR
Both formmulations are examples of the building block approach to search
strategy. This approach is especially suitable for this topic; since all info
tlon on-this topic is desired, the searcher need not be concerned with restricting
.the size nor with developing a high precisicn formulation. -Consequently,. the. ,-N
searcher is_more involved with expressing all aspects of the concept, word
variants and synonyms, rather than monitoring intermediary results
The Briefsearch would appear unnecessary here since:the topic is limited -to a
single facet and can be easily. searched using a few simple search statements.
arch.
(Reader: Keeping in' riind the stated information need of the requestor ., L6,retrieve all relevan items in,the ERIC ONTAPefile, which formulation would youchoosey
Be o valuating the results of the two formulations,.
up your own strategy and search the ONTAP file. Fotmulation 1f2 serves as an ex-,
ample for demonstrating the searching procedures of the ONTAP file. Whether you
construct your own formulation- or use one of the above examples, follow he -
planation below as you use the ONTAP file.
'After signing on to DIALOG, enter the BEGIN 201 command to start searching
the file. L _ _ review the explanation of the ONTAP file. Type ?ONTAP to view
the text.,
ONTAP(ONLINE TRAINING &-PpActicg) FILE'FRWIDE-S TEST OUE:TIONS RELEVNTCITATIONS FOR EACH OUESTIONTHE-ERIC FILE I: °: -THE -:OURCE-OF- THE- ANSWER
ET OUESTIONS AND ANSWER SETS APELABELLED AS TO COMPLEXITYE.G. --AN =S0AN=M04:IFINDDaJIENOTE_THE ANSWER-SETS':FOR SOD SIMPLEp-.--4TH-MEDiUM HID 3RD--DIFFICULTQUESTION;- AFTER-FINISHING-THFTEARCH THE PERFECT 'ANSWER SET FOR THATOUESTION- CANtTE SELECTED BY COMMANDS OF,THE FORM 'SELECT- -1=*=13.':, AND SELEQTED.ANZfWER SETS CAN BE COMBINED IN .AN-/AND'-PELATTONSHIR'WITH-YOUR FINAL..SEARCH:SET,TO FIND -OUT HOWANY-PELE%lANT CITATION=:YOU FOUND. A COMMAND-/EVAL COMPUTETHERECALL AND PRECISION 8F YOUR -SEARCH.-
EE ?TARSAM-FdP :T9i1PLE-sImpLE SEARCH..?TARSIM FOR LIST-,OFSfMPLE QUESTS.YT'TARMED-FOP .LISTOP: MEDIUM QUESTS.-.7TARDIP FOR LISTHOFDIFFICHLTQUESTIONSEVAL FOP OF PPEC.:'PEL.
At the conclusion of the above explanation are listed a number of
commands; the command,?TAPSAM, provides step-by-step directions for searching
the ONTAF file. ?TAFSIM, ?TAPMED, ?TAIDMED2, ?TAPDIF, TIAFDIF2, are lists of
ONTAP search questions and their corresponding "answer set." The command, ?EVAL,
gives the procedures for computing recall and precision scores. Using formulation
#2 as an example, we are ready to perform the search. Let's review all the steps
of the searching procedures by reading the directions triggered by the ?TAPSAM
command.
The _
sTARSAMTO USE ONTRP FILE, THE USER MUST BEGIN201:
1)SELECT QUESTION FROM PRINTED LIST ORDY STAPSIM, TTHAMED, STAADIF,.2) COMPLETE SEARCH IN YOUR' OWN WAYS)SELECT RN=CYOUR CHOSEN QUESTION - FORANSWER SET4i ANTI YOUR-FINAL SEA CH SET AND ANSWERSET5)ENTER EVRL COMMAND USING /EVAL(A,D,C)WHERE A =n OF CITATIONS IN SEARCH SET,
n OF CITATIONS IN ANSWER SET, C=-1z OFCITATIONS IN COMBINED SET
E.G: 1',SHOOSE QUESTION AN=S012)S DPAW(2W)PEASON(W)TEST (SET 1)
'SELECT AN= +_11 (SET 2)4)COMBINE AND 25:" ENTER -;'EVAL COMMAND,E.G. sEVAL(1,2,1)
TO IMPROVE YOUR RESULTS, TRY LOOKING ATOTHER PECORDS FOP SEARCHING IDEAS.
step in the directions recommends that w' select o. search topic
from the lists of simple, medium, and difficult questions. Since we have already
chosen parapsychology as a topic, let's move on to the second step, "complete
search in your own way." Formulation #2 is repeated below, but you can perform
the search using your own strategy.
59
Formulation #2
S FARAPSYCH?1 6 PARAPSYCH?
S ESP2 7 ESP
S EXTRA(W)SENSORY3 2 EXTRA(W)SENSORY
S EXTRASENSORY4 , 2 EXTRASENSORY,
S PSYCHIC?5 9 PSYCHIC?
C- 1-5/OR6 18 1-5/OR
At this point, it is necessary to find out whether parapsychology is a
simple, medium, or difficult topic so that we can compare our answer set with
ONTAF s "perfect answer set." Since there area number of search question lists,
let's print the list of simple questions first using the ?TAPSIM command.
?TAP: I M
I MPLE SEARCH TOPICS:301 DRAW-A-PERSON TESTSOa FID(INT. -FED. FOR DOC.)a,LIBRARYOR INFORMATION NETMOPKINGSO? 4-H C LUBS THEIR MEMBERS tlACT I V I TI ES
SO4 ' REV I S ION, ANGLO-AMERICAN CATAL . RULEStIE NAVAHO LANGUAGE TEXTBOOKS/GRAMMARS(MATEPIAL IV NAAHOI [IP USEFUL FOPTEACH I G NAVAHO, OP RFOUT NAVAHO L INGUIS-TICS)SO6 EDHCAT ION, -RI LANF;.AINCL. LIBRARY ,ACTIVITIES)OT' LITEPALY IN DEVELOPING COUNTRIES
PUBLISHED re THE INT=L. INSTITUTE FORADULT LITERACY METHOD: ANY PUPLIrATIONS FAY OR ABOUT THIS INSTITUTE)O 16 PER FACTOR TESTO9 PARAPSYCHOLOGY
Reviewing the list above, we find that parapsychology is simple search
question number 9. By selecting the accession number assigned to the topic, we
can obtain the number of postings i the "perfect answer set." The following
example demonstrates how to select the answer set.
60
Formulation #2 (
SELECT AN=S09 (or S AN=S09)
Comments
Selecting answer set to simplequestion 119
9 AN=S09 (Parapsychology) System response
The next step is combine our final answer set (search statement #6)
with the ONTAP answer net (search statement #7). In this way, we can find out
how many tens we retrieved are included in the ONTAP answer set. The result
of this operation is shown below, and indicates that we have retrieved all the
items in the ONTAP answer set in addition to some non-relevant material.
Formulation 2 (cont.)
6 AND 7
CoMMentS
Statements 6 and 7 are combined excluding
non-relevant material in our retrievedset
9 6 AND 7 System response
Although we can figure recall and precision scores by comparing the relevant
items listed by ERIC accession number at the conclusion of this discussion with
the final set (statement 116) of our search, ERIC ONTAP provides the /EVAL cepa-
bilityfor computing the scores. The directions are obtained by typing ?EVAL.
EVAL/EVAL0:4,B!C) COMPUTES THE RECALL ANDPRECISION nF ASEARCH. THE PARAMETERSAPE:
ATOTAL NUMPFP OF CITATIONS FOUNDIN YOUR FINAL ANSWER SET
ENUMPER OF CITATIONS IN ONTAPAN SET
C=NO. OF CITATIONS IN ANSWER SETFOUNDLBY USER ((RESULT OF
LOGICAL AND OF A AND B)RECALL=C/B PPECISION=C/A
PECALL TELLS YOU % OF RELEVANT CITATIONS IN THE FILE THAT YOU FOUNDPRECISION GIVES %.OF YOUP CITATIONSTHAT WEPE PELEVANT
After llowing the directions above, we can determine from reviewing the
online interaction that A 18 (total number of citations found in my final answer
set, statement #6), B = 9 (number of citations in ONTAF answer set, statement #7),61
and,C = 9 (number of citations in answer set found by COMBINing_statements
#6 and #7, statement 8). These values are inserted into the /EVAL formula
in order to obtain recall and precision percentages.
Formulation 12 (cont. ) Comments
/EVAL (18,9,9)
INPUT...EVAL(18,9,9)
Number of citations found = 18
Number of citations in answer set 9
Number of relevant citations found = 9
RECALL = 100% PRECISION = 50%
EVAL command with values A, B & Cinserted
System response to /EVAL formula
As mentioned previously, you can figure recall and precision ratios for
your performance by checking the results f your search with the list of ERIC
accession numbers wl the answer set to the parapsychology search topic.
Before moving on to the next search question, let's compare the results of the
two formulations for parapsychology.
Step 8: Evaluat n Final Results
As shown above, the second Tormulation free text searching Of
the ONTAP data base retrieves all the relevaot citation.: results of the
search are summarized below:
ulation #2
Number of citations found = 18
Precision
Number ofNumber of cite
I-
elzvant citations found 9 (100% relevant documentsons in answer et 9 retrieved)
Number of relevant citations found 9 (50% of retrieved docu=Number of citations'jound 18 ments are relevant)
Of` the eighteen citations retrieved by fo_ Ail- ion #2, nine are non-relevant;
.seven'of these areretrieved by the search term psychic used as a homograph
in the title or abstract. "Psychic energy," "psychic distance," and "psychic
development," serve as examples of phrases in which the term psychic occurs
62
but is deemed non-relevant to the search topic. In the remaining two non-
relevant items, ESP is used as an acronym for a program.
In contrast, formulation ;)1 retrieves six citations. Comparing its
answer =set with the -list of ERIC accession numbers at the conclusion of this
discussion, we can obtain recall and precision.
Fe -ulcti n
Number of citations found =
Recall = saber of relevant citations found 6 (67% relevant documeats
Precision =
Number of citations in answer set 9 retrieved)
Number of relevant citations found 5 (83% of retrieved docu-Number of citations found - 6 ments are relevant).
This formulation relies on the controlled vocabulary for search terms; in
this case, precision, the number of relevant items in the output, improved
when compared With the precision of the second formulation. The one non-relevant
item is retrieved because ESP occurs as an acronym in the identifier field. Of
the four relevant items which are not retrieved, two contain ESP as a term
in the title and two cone -in parapsychology as a term in the abstract. The
formulation fails to capture these four relevant documents because both the
,title and abstract fields, two sources of subject-rich terminology, are not
searched.
Since the requester is interested in retrieving all relevant information,
for-u #2 is preferred in this instance.
Relevant citations (cited by ERIC accession number):
EJ115774 EJ113605 EJ113604 EJ110046 EJ105855 E01102ED107626 ED104113 ED099252
EDUCATIONAL MANAGEMENT
ilk _AM. AM AMIL AMII
Search Topic #it Title IX, 1972 Federal Education Act Amend-
ments, and Women's Sports. (S AN=M10)
Search Objective: The expressed information need is to retrieve high recall
with a low number of non - relevant citations.
Step 3: Formul Basic Search _c i6 Plannin- Search _ate es
Three facets comprise the topic: sports, women, and Title IX. The diagram
below illustrates the three facets of the search. Where the three circles in-
tersect is the retrieved portion of the data base.
Subsearches for the three face the search assume the same strategy as
the sample search topic -p rapsychology. Just as all terms representing the
concept of parapsychology are searched and the results combined in an OR -ela-
tionship subseaithei for each facet of this topic will be performed in t
same way. All terms representing WOMEN, free text and/or assigned index terms
are SELECTed and then COMBINEd by an OR operator; likewise, SPORTS, and its var-
iant forms and synonymous terminology are joined by OR; a set representing TITLE
IX is formed similarly. The three subsearches are then COMBINEd by the AND oper-
ator, thus creating the final set.
65
66
Step the Searc
Women is a "lead-in" term in the ERIC Thesaurus directing the searcher
to females. From the list for females, descriptors indicating WOMEN are
gathered. The term sports like women, is not a descriptor but leads to
Athletics. There are no descriptors to represent Title IX or 1972 Federal
Education Act Amendments; consequently, determining variant forms of the Act
for searching the assigned index fields of the ERIC record can be done by re-
viewing the online dictionary for different forms which may occur in the iden-
tifier field, SELECTing from the list, and /or using these forms_to compose word
proximity phrases for free text searching.
Step 5: OrderingilnaRt_
Since the information seeker desires a high recall, with a minimum of non--
relevant citations in the output, the searcher must decide what approach to search
strategy would eliminate the majority of non-relevant citations from the output.
Selecting the most specific facet first approach may be a wise choice as the size
of the set representing the most specific facet, TITLE IX, may be small enough to-
terminate the search before applying any other facets to the strategy. At the
same time, the output can be monitored for loss of relevant material as search
facets are applied to the strategy. The Briefsearch could also be used initially
in order to estimate the size of the final set. The formulations below include
both approaches as well as the building block approach.
Step 6: Conce tualizin the Search as ln-u o the Retrieval S stemm
The Briefsearch, formulation #1 can be used to make a quick survey of the
ERIC file before a more complicated formulation is entered; rapid and efficient
in terms of online connect time, the Briefsearch can be formulated in a number of
ways.
Brief search
1 S WOMEN(C)ATHLETICS(C )TITLEMI
2 s WOMENIS(C)ATHLETICS(C)TITLEMIX
3 C 1 OR 2
Comments
Use of the (C) as the logical operatorAND eliminates additional steps forSELECT and COMBINE
The Briefsearch formulation retrieves three citations. At this point
searcher may call up the citations for online printing in order to check the
searching vocabulary used in the strategy with that of the free text and assigned
index fields of the document, or the searcher may continue searching with a more
complex formulation.
In formulation all subject-conveying fields are searched using free
text and word proximity. Since WOMEN, one of the facets of the topic, is a con-
cept which Occurs frequently as a population group in ERIC searches, a subsearch
is formulated with high recall as the goal; the search is saved for future use in
ERIC searches. By entering the .EXECUTE command, the Search Save for WOMEN is
performed and its results can be assimilated into the present search for Title IX
and Women's Sports. The formulation of the Search Save is given below. The appen-
dix contains 12 Search Save formulations, which are referred to in Section II
search topics.
Search facet: Women Comments
1 S DAUGHTER?
2 S FEMALE?
3 S FEMINI?
4 S GIRL?
5 S GRANDMOTHER?
6 S HOUSEWI?E?
7 S LADY?
B S LADIES?
9 S MOTHER?
10 S NUN
All subject-conveying fields are searchedin the Search Save for WOMEN. Highrecall is the primary goal of thesubsearch.
Included DESCRIPTORS and identifiersare FEMALES, FEMINISM, GIRLS CLUBS,HOUSEWIVES, MOTHER ATTITUDES, MOTHERS,NUN TEACHERS, NUNS, WOMEN PROFESSORS,WOMEN TEACHERS, WORKING WOMEN, WOMEN'SATHLETICS, WOMEN'S EDUCATION, WOMEN'SSTUDIES/Female, Feminist Press, Femi-nization, Masculinity Feminity Vari-able, Girl Guides Association, GirlScouts, Girl Scouts of America
67
Search facet con' t) Women Cvrffnen
11 S NUNS
12 S SISTER
13 S SISTERS
14 S WIFE
15 S WIVES
16 S WOM?N?
17 C 1-16 /OR
END/SAVESERIAL#1KLM
23MAY78 8:52:56 USER 4111$0.00 0.046 MRS FILE201 00 DESCRIPTORS$0.37 TYMNET$0.37 ESTIMATED TOTAL COST
Included DESCRIPTORS and identifiers areCamp Fire Girls, 3 multi-term MOTHERSDescriptors, 7 multi-term Mother iden-tifiers, Sisters, Wife, Wives, Women,11 multi-term Women and Women's iden-tifiers
Saved by the END/SAVE command, the formu-lation is assigned serial number 1KLM.The Search Save can be recalled andexecuted by the command .EXECUTE 1KLM
Procedures for exchanging Search Savesbetween DIALOG accounts is given inthe introduction to the appendix.
Free text searching is used'in formulation #2. The building block approach
characterizes this strategy as all three facets are developed in the formulation
and combined in the final step.
Formulation
1 =EXECUTE IKLM
2 S ATHLET?
3 S GYM?
Free text searching Common
4 S PHYSICAL EDUCATION
5 S SPORT?
6 C 2-5/OR
7 S EDUCATION(W)AMENDMENTS(F)1972
8 S EDUCATION AMENDMENT(F)1972
9 S TITLE(W)9
68
Search Save for WOMEN is processed
Included DESCRIPTORS/Identifiers:ATHLETES, ATHLETIC COACHES, ATHLETICEQUIPMENT, ATHLETIC FIELDS, ATHLETICPROGRAMS, INTRAMURAL ATHLETIC PRO-GRAMS, EXTRAMURAL ATHLETIC PROGRAMS,ATHLETICS, WOMEN'S ATHLETICS/AthleticAdministrators, Athletic Associations,Athletics for Handicapped, AthleticTrainers
GYMNASIUMS/Gymnastics, GymnasticsJudges
PHYSICAL EDUCATION, ADAPTED PHYSICALEDUCATION, PHYSICAL EDUCATION FACILI-TIES/ Physical Education and Recrea-tion, Physical Education Majors,Physical Education Teachers
SPORTSMANSHIP/Sport Ambassadors, Sportfor All, Sporting Goods, Sports, 6multi-term Sports Identifiers
10 S TITLE(W)IX
11 C 7-10/OR
12 C 1 AND 6 AND 11
Title IX
All three facets are joined together inthis final statement
Working with the most specific facet first approach allows the searcher to
inate the search when the final set shrinks to a size satisfactory to the
search objective. TITLE IX is selected as the most cpeci ic facet since it is
a named object rather than a concept like WOMEN or SPORTS. In order to determine
the different ways the phrase can be represented, the searcher uses the EXPAND
capability, SELECTs from the display, and formulates word proximit dhrases for
free text searching. The EXPAND display for the phrase TITLE IX/ID follows:
PEF INDE-TEPM TYPE ITEMS PTEl TITLE I-----------------
TITLE I HER1
E3 TITLE I PPESCHOOLFFOGFAM-: 1
E4 TITLE IIIE5 TITLE 1
E6 -TITLE L: = 12E7 TITLE OF THE EDUCATIC
H AMENDMENT:. ACT 0FS 1
E9 TITLE VII-,
F rom the above display, the searcher can either select the desired items
on to another EXPAND display for the Education Amendments Act of 1972.
E3 EDUCATION 221531E4 EDUCATION ACT OF 1972 1E5 EDUCATION AMENDEMENTS
OF 1974-1
E6 -EDUCATION AMENDMENTS-E7 EDUCATION AMENDMENTS OF
E :3 EDUCATION AMENDMENTS OF1972 TITLE
ELU AMENDMENTS OF1974-- -
Elf) EDUCATION. AMENDMENTS
Ell EDUCATION AMENDMENTS1972 TITLE IX---------
Ela EDUCATION AMENDMENTS1
El3 EDUCATION AND ECSTASY--
69
After viewing the EXPAND displays, the searcher can express the TITLE IX
facet in word proximity phrases.
Formulation
1 S TITLE(W)IX
2 S TITLE(W)9
3 S EDUCATION __ESDNENTS(F)1972
4 S EDUCATION(W)AMENDMENT(F)1972
5 C 1-4/OR
6 S ATHLET?
7 S SPORT?
8 C 5 AND (6 OR 7)
Comments
At this point, 39 citations are re-trieved. Rn:her than end thesearch, one Jf the other facetsis applied to the formulation
The combination of the two facetsTITLE IX and SPORTS results in 8
:r
citations. The search could beended or the third facet applied
In the above formulation, the application of the second facet to the first
reduces the postings from 39 to 8. The searcher could develop the third facet
but, in this case, has terminated the search rather than continue the online
interaction; however, this choice may affect the precision score as all facets
representing the topic have not been exhausted in th- strategy.
(Reader: In the light of the search objective, to retrieve high recallwith a low number of non-relevant citations, which strategy would youchoose?
Step 8: Evaluating Final Results
Before figuring recall and precision - scores, we, must find out what accession
number has been assigned to the search in the ONTAP file. Reviewing the fist of
all simple topics in the discussion of the parapsychology query discloses that
this topic on Title IX and Women's Sports is not among the simple topics. The
list of medium topics may be obtained by the command ?TAPMED.
70
-TAPMEDMEDIUM SEARCH TOPICS:MO1 DIPECT CHAPGING TO USERS' FOP PEFEREECE CUPRENT AMAPENESS SEPVICE OF
LIBPARIES OR OTHER INFO SERVICE_PHILOSOPHY,POLICY,PPACTICE,- FEECHARGES:ANY TYPE OF LIPPARY:ANYTYPE OF REFERENCE SERVICE;NOT INTEPESTED IN FFEE '3EPvIcE
ri']3 FEDEFAL AID TO -DAY CAFE CENTEPS OFSERVICES (INCLUDING HISTOPY,PHILOSOPHY,AFGUMENTS FPO CON,EXPEPIENCES,FUNDING,EVALUATION,PAPENTINVOLVEMENT AND ATTITUDES) ONLYH. S. CENTERS OP SERVICES
MO.=: JEAN PIAGET'S THEORIES, AMP THOUGHTPPOCESSES OP LANGUAGE DEVELOPMENTOF CHILDPEN,BUT LIMITED TO PUBLI
_ CATION AVAILABLE FPOM FP1C/EDPSFOP MORE, ENTEF TTAPMED1
7TAPMED1MEDIUMSEAPCH TOPICS:MOO L1BPARIES LIBPARIAN MIDDLE
EAST,(FXCLUDE AFPICAN BUT NOTEGYPT) . INCLUDE ALL TYPEF OF LIBRARIES INFORMATION.CENTERS
MO5 LIBRARY SERVICE TO PHYSICALLYHANDICAPPED (NOT MENTALLY OFLANGUAGEHANDICAPPED)
'ti 6 EFFECTS OF TV VIOLENCE OM CHILDRENMO7 DFUG ABUSEIINCLUDING ALCOHOL, AMONG
STUDENTS OF ELEMENTAPY OR SECONDARYSCHOOLSIGPADES K-12,:INCLUDINGCHOOL EDUCATION PROGRAMS AND
SOCIOLOGICAL STUDIE' )mos SCHOOL BUSING RACIAL INTEGRATIONMil' PECPEATIONAL LITE OF FOPEST LANDSM10 TITLE 9,1972 FEDEFAL EDUCATION ACT
AMENDMENTS,AND WOMEN'S SPORTSti WHITE FLIGHT TO THE SUBURBS
Scanning the above list, one finds that the search question has been
assigned accession number M10. To determine how many citations are relevant
to the search topic, select the accession number, SELECT AN = M10. Using
the EVAL procedure, discussed in the sample search, the results of the three
formulations are evaluated. The Briefsearch, formulation #I, fared well, re,trieving three citations, two of which are relevant; in this case, it functions
as a survey of the file, retrieving 50% sf the relevant items in the file.
71
Brief search
Number of citations found m 3
Recall =
Precision -
Number of relevant citations foundNumber of citations in answer set
Number of relevant citations found 2
Number of citations found 3
(50% relevant documentsretrieved)
(66% of retrieved docu-ments are relevant)
Formulation 4l3 in which the most specific facet first approach is demonstfated
retrieves all the relevant items but also as many non-relevant items,
Formulation 1/3
Number of
Recall
citations found 8
Precision
Number of relevant_umber of citations
ations found 4 (100% relevant documentsin answer se 4 retrieved)
Number of relevant citations found 4 (50% of retrieved docu-Number of citations found 8 ments are relevant)
Formulation //2 retrieves all the relevant items, no non-relevant it
resulting in 100% recall and 100% precision. As the information need is stated,
the requestor would be satisfied with a high recall search as long as few non-
relevant citations are included in the output. Both formulations 4l2 and #3 pe
form well in terms of recall; however, the precision of formulation #3 in
which the most specifid facet first approach is used, is lower than formulation #2,
the building block approach. Formulation 13 is more efficient and rapid in terms
of online connect time since only two facets are developed in the strategy. Thus,
either formulation #2 or #3 is preferred for the search topic.
Relevant Citations:
EJ119209 EJ105832 EJ105671 ED110452
72
CAREER EDUCATIONF
,-h topic #2: 4-H Clubs, their members and activities.
(S AN -S03)
Step 3: Formulati e Ba Piaunin Search_Strate_i
A single concept is present in the topic: 4-H Clubs. Consideration must
be given when constructing the formulation<to include variant forms and spell-
ings of the Club.
Step 4; C the Search Terms
No descriptors are listed in the ERIC Thesaurus which directly refer to
4-H Clubs. Search formulations are constructed using the word proximity feature
in order to retrieve different spellings of the Club-.
Step 5; Orci_LaThpitut
No output specifications are requested.
Step 6: Conce ualizin the Search as In.ut tc the Retrieval System
If a high precision search required, you could search only the identifier
field as in the following formulation:
FOrmula n #1
1 S 4(108/ID
S FOUR(W
3 C 1 OR 2
HRH
Comments
Using the word proximity feature, additional citations are retrieved incases where another word. e.g. 4 HClubs, follows the identifier phraSe.Without the proximity feature, onlytwo citations are retrieved in com°'parison to 14 citations
The following search formulation is a high recall formulation in which all
subject-conveying fields are searched:
Formulation #2
1 S 4(W)H
2 S FOUR(W)11
3 C 1 OR 2
Comments
All subject-conveying:fieldsin this formulation
are searched
-_-Step 7: EValustiag_12Iliminary-Results
In formulation #1 fourteen citations
formulation. At this point the number of
both cases that
ties" aspect of
cific facet firs
h
a- retrieved, 21 in the second
citations is sufficiently small in
s felt to be unnecessary to impose the "member" or "activi-,
search. This is an example of starting with the most ape
Step 8: 'Evaluating Final Results
Searching the identifier field in the first fo
precision but some relevant items are
than fOrmulation #2
lation results in high
missed resulting in a lower recall score
in which all subject rich fields are searched. The recall
and precision scores forhoth formulations are summarized below:
Formulation #1
Number
Recall
of citations found = 14
Number of relevant citations foundNumber o ,citations in answer set
Precision -Number of relevant citations found
Number of citations found
Formulation #2
Numbe
Recal
Precision
ations found 21
Number of relevant citations foundNumber of ations in answer set
0
Number of relevant citations foundNumber of citations found
Relevant citations:
EJ120500 EJ117540 EJ113715
ED109351 ED109205 ED105103
ED102469 ED100323 =ED095289
14 (82% relevant Aoc-___ents17 retrieved)
14 (100% of retrieved docu-14 meets are relevant)
17 (100% relevant documents17 retrieved)
17 (85% of retrieved 'docu-21 meats are relevant)
EJ109322 EJ107077 EJ107045 ED109352
ED105102 ED105101 ED105100 ED104895
COUNSELING AND PERSONNELSERVICES
Search Topic #3: 16 Personality Factor Test.
Search Objective: The stated information need is to retrieve
a maximum number of relevant citations using
the least amount of online connect time. (S AN=S08)
Step 3: Fo nlatin Basic Search Lo c Plannin Search S
A single facet is present in the search topic: 16 Personality Factor Test.
The main difficulty with this search is the different spellings of this test.
Step 4: Co lin the Search Terms
Since no descriptors are listed in the ERIC Thesaurus, you could EXPAND the
numeric character 16, as an identifier, SELECT from the online dictionary; then
EXPAND 16 spelled-out and SELECT from that portion of the online dictionary for
identifiers.
Step 5: Ordering Output
Keeping in mind the search objective, the searcher should strive to reduce
online connect time by preparing before the online interaction, cu tailing-ex-
cessive typing in order to minimize typing errors, and selecting a suitable
approach to search strategy. The Briefsearch may be favored because of
efficiency; variant spellings of the test could be represented by single search
statements using word proximity.
Step 6: Conce he Search as In.ut to the Retr -v S stem
The following formulation takes into account different spellingp of the test
garnered from online dictionary displays:
76
_Formulation t1
1 SN16PF
2 S 16(W)PF
3 S 16(W)P.(14)F.
4 S 16(W)PERSONALITY(W)FACTOR
5 S 16(MPERSONALWY(W)FACTORS
6 S SIXTEEN(W)PERSONALITY(W)FACTOR
7 S SIXTEEN(W)PERSONALITY(W)FACTORS
8 C 177/0R
Comments
Free text searching is used exclu-ively in this formulation.
Includes Identifiers: 16PF, 16 P. F.,116 Personality Factor Question-`naire, 16 Personality Factor Test,Sixteen Personality Factor Ques-tionnaire
As an abbreviated form of the above formulation, the following Briefsearch
contains all the elements of the longer search except search statements 2 and 3.
Since two possible ways of spelling the .test are not included in the formulation,
a lower recall score may be expected.
Brief search
1 S 16PF
2 S 16(W)PERSONALITY.
3 S SIXTEEN(W)PERSONALITY
4 C 1-3/OR
Comments
Includes all of the Identifiersnamed above except 16 PF and16 P. F.
(Reader: Keeping in mind that the information seeker is concerned
with online connece'time, which formulation would you choose?
Step 8: Evaluating Final Results
In order to find out what citations are relevant to the topic, the searcher
must review the lists of simple,' medium, and difficult topics to found out the
accession number of the search. From the list of simple topics, we find that
the topic is assigned number simple 8 008Y Which we then SELECT (S AN=S08)
in order to obtain the number of relevant items. Surprisingly, both formulations
fared the sake, retrieving all 12 relevant items and one non-relevant citation,
Recall and precision scores for both formulations are summarized below:
Formulation ll1 And Brief-search
Number of citations found = 13
Recall -=
Precision
Number of relevant citations found 12 (100% relevant docu]Number of citations in answer
Numbe
set , 12 retrieved)
elevant citations nd 12 (92% of retrieved docu-13 meats are relevant)Number of citations found,
Since the online dictionary is called up in formulation is time-
consu g,-thus costly. In a trial run of the first formulation in which
tour EXPAND commands are used, the search takes .125 (71/2 minutes) to perform.
In comparison, the Briefsearch takes .25 (11/2 minute and, in this case,
retrieves the same set of documents. Because of its efficiency, the Brief-
search is preferred for search topic
Relevant citations:
EJ120592 EJ117340 EJ117192 EJ115601 EJ115600 EJ112211
EJ108963 ED103575 ED103574 ED102449 ED101228 ED095465
77
READING ANDCOMMUNICATION SKILLS
Search Topic 114: Evaluation of Primary School (grade 1(-3)
English reading programs or reading materials and techniques
(not evaluation of specific reading teats or-instructors, no stu-
dent test scores when not part of evaluation of reading program,
and not just criteria or standards for evaluation). (S AN=D03)
Search Objective: To retrieve as much relevant material as possible with a
minimum of non-relevant citations in the output, Only
publications available from ERRS are desired;
Step 3: Formulating Basic Search Lo =ic Plannin Search Strategies
The search topic can be broken down into three facets: reading, primary
school, and evaluation. Illustrated in the following diagram, the search topic
is separated into its three facets which intersect to form the portion of the-
data base intended for retrieval.
PRIMARY EDUCATION
Representing those aspects of the query which are not wanted, e.g., evaluation
of reading tests, evaluation of instructors, may eliminate relevant material from
the output. For example, a set created by the statement "EVALUATION AND READING"
may have documents dealing with the evaluation of reading teachers as well as how
teachers evaluate reading materials, so that a set created_ by the statement
79
'READING AND (EVALUATION NOT TEACHERS)" would exclude valuable material.
.'7Consequently, the search formulations constructed below will not employ the
Boolean NOT operator.
Step. 4: Compiling the Search Ter
All facets of the'search are well represented by terms froMthe ERIC
Thesaurus. In _'act, entering READING as a free teksearch term retrieves
occurrences of the term in 58 multi-term READING descriptors, 52 multi-term
READING identifiers and all occurrences in fre fields, e.g. title, abstract
or corporate'source. EVALUATION occurs in 26 ulti-term descriptors, and in
10 multi-term identifiers as well as in free text fields. The abundance of sub-
ject terminology acts detrimentally as the search is made difficult by the volume
Of literature in the ERIC file on reading and on evaluation.
Step 5: Ordering Output
In order to restrict the search to documents available from ERRS, the
LIMIT command can be applied at the beginning of the search. Applying the
LIMIT after the creation of the.first facet allows the searcher to monitor
preliminary search esults throughout the online interaction.
The ERIC file/ contains much material on reading and evaluation, so.
that the searcher Should select an approach to search,strategy which enables
constant evaluation of-preliminary search results. In this way, the searcher
ca ntroduce limiting criteria to the formulation as the search is in pro-
gr ss. For example, in the building block approach, the combination of
all three facets occurs as the final statement of the =a h' if the r
sul ant set is too large, the searcher has little recourse except to
"backtrack." or reformulate some of,the essential elements of the search.
On the other hand, in the successive fractions approach, the
Br,
search is.limited to the initial "bite" or partitio- of the file; subsequent
search facets continue to be applied to this partition so that the searcher
can terminate the search if necessary before exhausting all search facets.
and
The sample formulations provided in step 6 cover the successive fractions
building block approaches. Follow the development of the formulations
especially by reading the COMMENTS section, so that you are aware of the de-
cisions made by the searcher as the search proceeds.
Step 6 Conce.tualizin= the Search as In to the Retrieval S
Because the volume of literature on Reading is so large, you could formulate
a Briefsearch first in order to perform a survey of the file. This will give you
ideienf the possible size of the final set.
Brief search
S PRIMARY(W) EDUCATION (F)READING(F)EVALUATION
) EVALUATION2 S PRIMARY(W)GRADES(F)READING
3 C 1''OR 2
4 LIMIT 3/AVAIL
Comments
Twenty citations are retrievedin the Briefsearch giving anestimation of the volume ofoutput
Since the terms READING and EVALUATION"have such high posting counts, the
formulation might concentrate on the identifier, descriptor, and title fields in
order to eliminate the possibility of retrieving false drops due to casual mention
of the search terms Reading and Evaluation in the corporate source, abstract or
descriptive note fields. Using Search Save l/1 (see appendix) for representing
the facet of PRIMARY EDUCATION, the following strategy, an example of the build-
ing block approach, searches the assigned index fields and tiele for the READING
and EVALUATION facets and all free text fields for the PRIMARY EDUCATION facet.
81
Formulation 1/2
1 .EXECUTE (Search Save serialnumber for PrimaryEducation)
Comments
2 LIMIT 1 /AVAIL The LIMIT is applied to the setthe facet PRIMARY EDUCATION
as the second step in this formu-lation
3 S READING/DE,ID,TI Reading, as a search term, is re-stricted to occurrences in the ti-tle and assigned term fields.
4 S COMPARATIVE ANALYSIS/DE,ID,TI Statements 4-8 form the facetEVALUATION; searching is limitedto the title and assigned term
5 S COMPARATIVEMSTATISTICS/DE,ID,T1- fields.
6 S COMPARATIVE(W)TESTING/DE,ID,TI
7 S EVALUATION /DE,ID,TI
8 C 4-7/OR
9 C 2 AND AND 8
The high-recall search formulation employs Search Sava 111 for the PRIMARY
EDUCATION facet and Search Save 1/7 for the EVALUATION facet (see appendix).
Formulation 113 is an example of the successive fractions approach to search
strategy; as this search progresses, note what factors are:_applied to the for-
mulation in order to restrict volume of output..
Formulation 1/
Part I:
1 .EXECUTE (Search Save forPrimary Education)
2 LIMIT 1/AVAIL
3 S READER?
4 S READING
5 .C 2 AND (3 OR 4)
.EXECUTE (Search Save forgeneral EVALUATION)
Comments
Two facets of the search, READING andPRIMARY EDUCATION are joined by AND;the large number of citations (307)retrieved at this point makes it clearthat further reductions are necessary.The next step is to apply the generalSearch Save for EVALUATION to thesearch.
Formulat on (cone.
Part I (cont.):
7 G 5 AND
Comments
The COMBINing of the three facets re-sults in 250 citations., At this pointthe search could be concluded if thesearch objective is high recall.
Part II:
8 .EXECUTE (Search Save for Since the output of the high recallrestricted formulation is large, the formulationEVALUATION) is reconstructed using Search Save #8
for EVALUATION. Search terms arelimited to descriptors whose occurrencesare retrieved from the assigned indexand title t!'
Here, the volume of output is reducedto a total of 121 citations
Another,apprcach to restricting theformulation i3 to limit the READINGfacet to oorcurrences of Reading intitle, identifier and descriptorfields.
At'this point, 196 citations are re-trieved, and it is decided to continueintroducing limiting criteria to thesearch formulation.
9 C 8 AND 5
S READING DE,ID,TI
Part III:
12 S ACHIEVE?
-- 13 S APPRAIS?
14 S ASSESS?
15 S GOMPAR?
16 S EVALUAT?
17 C 12-16/OR
18 C 2 AND 10 AND 17
Instead of using the comprehensiveEVALUATION Search Saves, a formulationcontaining a few search terms' from theEVALUATION Search Saves is constructed.
Statements 12-17 make up A subsearzhfor EVALUATION
137 citations are retrieved using theabbreviated subsearch for EVALUATION,Search Save for PRIMARY EDUCATION, andthe restricted READING set.
(Reader: Given the many possible approaches to formulating thesearch strategy, which formulation would you choose? Keep inMind that the information seeker desires a minimum of non-relevant citations in the output.
Step 8: Evaluating Final Results
The list of difficult questions in theONTAP file is given below and must
be checked for ONTAP accession numbers. In this way, one can SELECT the accession
number (S AN=D03) in order t. find out how many citations are relevant to this topic.
?THADIFDOI TRAINING FOP SUPERVISION MANAGE
MENT IN LIPPORIES INFORMATIONCENTEASINCL. NEED FOP TRAINING! DESCAIPTIONS OF TPAINING PROGRAMS ORMATEPIALS,TRAINING OF STUDENTS ANDPROFESSIONAL WORKING LIBPARIANS;'ACADEMIC OF ONTHEJOE TRAINING
D0 2 AUDIOVISUAL AIDS FOP,OPIENTATIONOP INSTRUCTION OFLIEPHAY USERS-EVALUATION OF PFIMqPY SCHOOLS GRADE
ENGLISH READING PROGRAMS OFREADING MATERIALS TECHNIQUESaiOTEVALUATION (IF SPECIFIC REAPING TESTOF INSTAUCTOPS,NOT 'UUDENT TESTSCORES ''HEN NOT PRAT OF EVALUATIONOF READING PAOGAAMO, MOT JITIT THECRITERIA OP TAN[PDT FOP EVALUATION)LIMIT TO FUES.AVAIL.EPIC/EDRS
DO4 roPrIFIL SCIENCE FlIUGATION PAOGAAMSIN U4IVERSITIE7i SECONDARY GPADES9-12 IN SOVIET UNTONkALL SCIENCESINCL.MATWENGIN.STAADIF1 FOP MORE
? ?TAADIF1D05 VOCATIONAL EDUCATION OF AMERICAN
INDIABI(HISTOPYDATA! PROGARMS-TOPROVIDE THIS EDUCATION :BUT NOTTARTNING MATEPIALS TO BE USED
DOE, EVALUATION OF BILINGUAL ELEMENTARYELEMENTAPY (KA)- -ECONDARY (9-12)SCHOOL PROGRAMS- ORTECHNIPUEs,: FOR'SPANISH EN5LTSH LANGUAGES:
DO7 FINANCIAL COS:T.Tri INSTITUTIONS- OFHIGHER EP= TO IMPLEMENT LEGILATEDAFFIRMATIVE ACTION APOGPANST
DO8 EVALUATION OF INDEXING CATALOGING('WORK1METHODS.PRODUCTSLAN1URGES)TOINCLUDE REPRESENTATION STOARGE OFTHE INDEX INFO. 'INDEXING ' INCLUDESALL FOAMS OF TEXT SEARCHING,I.E.INDEXING EY TEXT WORDS./EVALUATION''IS TIME ('CO-.ST, ERROR PAT 1--ALL/
RELEVANCE! AND EVAL.CRITEPIA) DEXING INCLUDES CITATION INDEXING!TITLE wORD INDEXING! OR AUTO INDEX
DO9 COLLECTIVE BARGAINING IN LIEPAPIESOF INSTITUTIONS OE HIGHER EDUCATION
84
Forty -six documents are judged as being relevant to the information -need.
In formulation #1, a minimum of'online searching effort is expended. The re-
sults are as follows:
Formulation 111. (Briefsearch)
Number of citations found = 20
Number of elevant citations foundRecall =citations in answer set
Precision =
Number
Number of relevant citations foundNumber of ations found
15 (33% relevant documents46 retrieved)
15 (75% of retrieved docu-20 nents are relevant
The Briefsearch perfo_ s well in terms of precision, capturing fifteen rele-,
vent documents out of the twenty retrieved items. Tete information seeker, pre-
rented with the results of this search does
vent-material although 26 relevar
formulation #3'which is
not have .to sort through much non-rele,.
items have been missed. On the other band,
constructed for high recall retrieves a high percentage of
relevant items at the expense of re icving a high volume of non-relevant item
Recall and precision ratios for three parts of the sgarch are listed below:
Formula i. Using the general EVALUATION
Number of citations found = 250
Number of relevant citations foundNumber of citations in answer set
Recall
Precision =
Formulation
Number of
Recall
Number of relevant citations foundtions foundNumber
Search - Save)
46 (100% relevant documents46 retrieved)
(Using the restricted EV
citations found 196
Number of relevant citations foundNumber of citations'in answer set
of'relevant citations oundNumber of citations found
NumbePrecision =
Formulationj122IrIIT, (Using the abbrev
Number of citations found = 137
Recall =
Precision
Number of relevant ci atNumber of citations in a
Number of relevant
Number of cit
46 (18% of retrieved docu-250 ments are relevant)
ATION Search Save
42 (90% relevant documents46 'retrieved)
42 (22% of retrieved doc-196 uments are relevant)
ted search for EVALUATION)
39 (85% relevant documener set 46 retrieved)
citations found 39 (28% of retrieved.docu--ns found 137 ments are relevant)
86
As formhlation 113 proceeds, additional restricting factors are applied to
the search. In part 1 the three facets are made up of two atiin2 Search Saves
for EVALUATION and PRIMARY EDUCATION, and the occurrences of READING or READER
in all subject-conveying fields. Here, all 46 relevant citations are retrieved;
however, over 200 retrieved citations are not relevant to the information need.
In part II, the general EVALUATION Search Save and Search Save for PRIMARY
EDUCATION are COMBINEd in an AND relationship with a restricted READING set.
These limiting factors result in the loss of four relevant items; however, 50
non-relevant items are not included in the final set.
Additional restricting factors applied in part II of ormulation #3 result
in improved precision in exchange for a lower recall. The third formulation, an
example of the successive fractions approach to search strategy, performs well,
in this case, for a high recall search in which all relevant citations are de-
sired; but the search objective as expressed by the information seeicer is,00t
satisfied since over two-thirds of the output in all three portions of the search*
centains non-relevant material.'
Search formulation 112, made up of restricted sets for EVALUATION and READING,
works out surprisingly well in terms of precision and recall values as shown
below:
Formulation #2
Number of citations found = 52
Recall
Precision
In view
Number of relevant citations found 30 (65% relevant documentsNumber of citations in answer set 46 retrieved)
Number of relevant citations found 30 (58% of retrieved docu-Number of citations found 52 ments are relevant)
information -neede expressed by the information seeker, the
results of the Briefsearch or formulation #2May be satisfactory.
doe are retrieved in formulation #2 than the Briefsearch at
drawing out some non-relev
More relevant
the expense of
aterial; however, the Briefsearch serves as a high
pLecisioni low cost fo ul t on.
Reader: Since there are a number of possibilities present in
choosing among formulations, which would you
reading the evaluation?
Relevant citations:
select after
ED110543 ED110511 ED110510 ED110171 ED110170 ED1101i)8 ED109934
ED109686 ED109674 ED109638 ED109632 ED109631 ED109592 ED109150ED108791. _ED108377 ED108201 ED108197 ED108196 ED108192 ED107372
ED107071 ED106750 ED106400 ED105440 ED105012 ED104999 ED104994ED104925 sED104919 ED104527 ED103008 ED103004 ED100176 ED100175ED099822 ED099820 ED099812 ED098494 ED098346 ED097645 ED097635
ED097104 ED096781 ED096780 ED096033
87
SOCIAL STUDIES
=
Search Topic 5: Use of school busing to
tegration. (S AN=M08)
Step Formula Basic Search to io,- Plannin
racial
o facets comprise thes arch topic: integration and busing. The concept,
of school could form another facet; however, since:the type of school is not
specified in the query, the facet of school will not be applied to the search _
results- unless a large volume of "output is retrieved in the-formulation below.
Step 4: Compiling the Search Terms
The INTEGRATION aspect can be described by a number of ERIC descriptors, in
comparison to BUSING whidh is-only represent djly three descriptors, bus
transportation, student transportation, and school buses. Formulations con-
structed for free text searching-can include variant forms of the busing and
desegregation since neither term is used singly as a descriptor.
Step 5: GideiL.n_hpit7ut
No output specifications are stated in the information need. .Thus the
discussion will cover three different Search strategies in order to illustrate
how each of these approaches is employed to achieve one of three objectives:
high.recal4 high precision,
Step 6: nee ualizin the Sea
low cost.
h as
Whether to
ieval S0
vey_ the file or gather additional searching vocabulary,' the
ieves a high percentage of relevant items at a0_owBrie_ arch given below
cost.
Fo on
1 S STUDENTMTRANSPORTATION C INTEGP TON
S BUSING(F)DESEGREGATION
3 C 1 OR 2
Co
After COMBINing the results ©fthe two sets, 23 citations areretrieved. At this point, itis not felt necessary to addthe third facet SCHOOL to theformulation and the search isterminated.
As a result of evaluating t :e Briefsearch, it is decided that the third
facet of SCHOOL will not be included in the high recall formulation, an example
f the building block approach, which
Formulation #2
follows:
Comments
1
2
3
4
5
6
7
B
10
11
12-
13
14
15
S
S
S
S
S
S
S
S
C
S
S
S
S
C
C
BUSBUSES.BUSINGBUSEDBUSSESBUSSINGBUSSED
STUDENT(C)TRANSPORTATION1-8/ORRACIAL?INTEGRAT?DESEGREGAT?SEGREGAT?10-13/OR14 AND 9
Using truncation to retrieve the manypossible forms of BUS would probablyresult in a large number of false drops.Each form is typed in separately ratherthan entering.BUS?
Included DESCRIPTORS/Identifiers: BUSTRANSPORTATION, SCHOOL BUSES/Bus Drivers,Bus Driver Training.
Included DESCRIPTORS/identifiers: 22INTEGRATED & INTEGRATION DESCRIPTORS,SEGREGATED PUBLIC FACILITIES, 7-SEGRE-GATION DESCRIPTORS, SEGREGATIONIST OR-GANIZATIONS/Desegregation Aid, Desegre-gation Advisory Project
Fornielation 113 is an example of the lowest postings facet first approach.
Expected to produce high precision,- the strategy involves the development and
eValuatin of the facet BUSING before the application of the second facet
INTEGRATION.
Formulation t3 Comments
1
2
3
S STUDENT TRANSPORTATION
S BUS TRANSPORTATION
S SCHOOL BUSES
Using ERIC descriptors to create a setof documents for the facet BUSING shouldresult in high precision since only theassigned index fields are scanned.
4 C 1-3/OR At this point, 63 citations are retrieved.The searcher could end the search now,but decides to add the second facetINTEGRATION which is constructed usingfree text searching.
5 S INTEGRAT?
6 S DEGREGAT?
7 S SEGREGAT?
8 C 5-7/OR
9 C 4 AND 8The BUSING facet is reduced to only thosedocuments containing search terms forINTEGRATION in any of the 7 subject-con-veying fields.
Step 8: Evalua in- Final Results
The Briefsearch is remarkably accurate in terms of retrieving relevant:documents
as reflected by high precision. However, all 42 of the relevant items in the file
are not retrieved by this formulation. The Briefsearch is favored if high pre-
cision and low cost are search objectives.
Briefsearch
Number of citations found = 23
Recall = Number of relc ant citations found 21 (50% relevant documents
Precision
umber of citations in answer set 42 retrieved)
Number of relevant citations found 21_ (90% of retrieved docu-Number of citations found 23 ments are relevant)
On the other hand, if high recall is desired by the requester, formulation #2,
the building block approach, in which free text searching is used may be preferred.
Capturing all the relevant items, the formulation also results in retrieving a
high volume of non-relevant citations.
Formulation
Number of citations found = 63
Recall =
Precision
Number of relevant citations foundNumber of citations in answer set
42 (100% relevant documents42 retrieved)
Number of relevant citations found 42 (66% of retrieved docu-63 ments are relevant)NUMher of citations found
Formulation 43 in which high precision is intended achieves that objective.
Developing the BUSING facet by descriptor terms alone is not suitable, in this
case, to effec t high recall also.
Formulation 43
Number of citations found = 29
Recall =
Precision ----
Number of relevant citations foundNumber of citations in answer set
*26 (607 relevant documents42 retrieved)
Number of relevant citations found 26 (90% of retrieved doc-Number of citations found
There is a "trade,
29 uments are relevant)
" in each of the formulations. For example, formulation
3 achieves high r carl at the risk of retrieving much non-relevant material. At
the expense of recall, the Briefsearch is a low cost, high precision formulation.
Compare your search `armulations with those given. Check to see whether you have
compromised in order tI achieve high recall, or vice versa.
Relevant citations;
Ed 121907 Ed 121906 EJ 121889 EJ 121888 El 121886 EJ 121885Ed 120204 Ed'115861 Ed 115858 Ed 115853 Ed 115849 EJ 115809Ed 115252 Ed 114427 rEd 113526 Ed 113504 Ed 111820 Ed 110130EJ 110129 Ed 110036 Ed 108492 Ed 108055 Ed 107317 EJ 107301Ed 106768 EJ 105109 Ed 103409 Ed ].03405 Ed 103347 ED 110551ED 110533 ED, 109745 ED 106425 ED 1.06397 ED 105554 ED 103546ED 103288 ED 102279 ED 102242 ED 101045 ED 101018 ED 099258
92
LANGUAGES AND LI lGUI STICS
Search topic Navaho language textbooks or grammars
(material in Navaho or useful for teaching Navaho, or about,
Navaho linguistics). AN S05)
Search Objective: The information seeker is interested in gathering as much
information as possible on this topic at the lowest cost.
Step 3: Formulatin Basic Search Lo plannint Search S-ra
Two concepts are present in the topic: Navaho and language. The diagram
below illustrates the search logic; the shaded section represents the portion of
the data base intended for retrieval.
Step 4: Com i in the Search Tc
There is sufficient terminology for expressing the facet LANGUAGE in assigned
index terms. The NAVAHO facet can be represented by=a single descriptor, no that
related term; such as American Indian Languages or American Indian Culture could
be introduced to the formulation; however, these terms may retrieve items treating
American Indian Languages as well as the[ Navaho language.
Step 5: Ordering Output
Selecting an approach to search e y should take into consideration the
search objectives: high.recall and low coat. The most specific facet fir
approach, ,since-it permits the searc=her to terminate the search before introducing
all facets to the formulation, is prefer re as well as the Briefse Developing9.3
a high precision formulation is also necessary as the cost of the search
involves offline prints as well as online connect time.
Step 6: Conce ualizin the Search as Input to the Retrieval System
A Briefsearch is constructed by taking variant spellings of Navaho and
combining them with the search term Language using the AND operator. The (C)
limiter could be used instead of the AND operator, however, it cannot be used in
conjunction with truncation.
Briefsearch
1 S NAVA ?O? ?
2 S LANGUAGE?
3 C 1 AND 2
Comments
Different spellings of NAVAHO are easily re-trieved - employing DIALOG's truncation feature;since the root of the term is not common tomany words, few false drops should occur.
Since 49 citations are retrieved after de-veloping the most specific facet, NAVAHO,the searcher decides to apply, the second ,
facet rather than risk retrieving much non-relevant material.
The Briefsearch e ieves 20 citations,; the searcher could end the search
at this point or continue developing these statements into a more complex forma-
iation, FOrmulation #2 incorporates terms taken from the search topic into the
LANGUAGE facet. A greater volume of output can be expected as citations are
trieved which contain not only Language, but Textbooks or Books, or Grammars, etc.
Formulation 112
4 S BOOK?
5 S GRAMMAR?
6 5 LINGUISTIC?
7. S TEXTBOOK?
8 C 2 OR 4 OR 5 OR 6 OR 7,
9 C 1 AND 8
Comments
Formulation 1/2 uses the groundworkprovided by the Briefsearch; in thisway, the. searcher does not have tobacktrack to create sets for NAVAHPand LAV;UAGE.
Step 8:
(Reader: After executing your own search and reviewingthe sample formulations, which formulation would youselect? Keep in mind that the requestor is concernedabout cost.
Ev Final Result
,The Briefsearch manages to capture eight of the nine relevant items in the
ONTAP file. Only three search statements are entered in this formulation keeping
online connect time at a minimum.
Briefsearch-
Number of citations found = 20'
..umber of elev. vitaRecall
Precision
ns foundNumber u= citations in answer set
8 (89% relevant documents9 retrieved)
Number of relevant citations found 8 (40% of retrieved doc-20 uments are relevant)umber of citations found
Just as the Briefsearch retrieves only eight of the nine relevant documents,
the second formulation fared the same; however, three additional non-relevant
citations are picked up as a result of including more search terms into the devel-
opment of the LANGUAGE facet.
Formulation #2
Number of _ ions found = 23
Recall = Number of relevant citations foundNumber of citations in answer set
Precision
(89% relevant documentsretrieved)
Number of relevant- citations found 8 (35% of retrieved docu-Number of citations found 23 meets are relevdnt)
Both formulations fail to retrieve the same item (ED 104168) judged as being
relevant to the topic; curious to find out why the document is missed,, we
printed it in full format. No portion of the record mentions specifically the
Navaho language; it deals generally with North American Indian languages, so
that its relevance to the topic may be questioned.
Relevant citations:
EJ 111077 EJ 111075 ED 108801 ED 104168 ED 102820 ED 102585ED 101549 ED 101152 ED 100136
INFORMATION RESOURCES
AMTNAdia Ai. Ali Am.Search topic Revision o f the Angl 'm an Cataloging
Step 3:
Rules. CS ANS04)
in- Basic Search Log Search Strategies
Two concepts are present in the topic: REVISION and ANGLO -A RICAN CATALOGING
RULES. The shaded section in the diagram below indicates the portion of the data
base intended for retrieval.
REVISION
ANGLO AMERICANCATALOGING RULES
Step 4: Compiling; Search Terms
Anglo American Cataloging Rules, AACR or revision are not included as
descriptors in the ERIC Th aurus __a much of the formulation will consist of
free text searching. A high precision search would be restricted to the ideati
fier and title fields.
Step 5: Orderiit_ut
No output specifications are identified in the query. The formulations
give examples of the Briefsearch, building block approach, and a high recall
formulation in which the identifier and title fields are searched.
Step 6: Conceptualicrng the Search as Input to the Retrieval S
Formulation #1 serves as an example of the building block Approach. All
subject - conveying portions of the ERIC record are searched for occurrences of
the search terms as free text searching is employed.
Formulation
1 S AACR
2 S ANGLO CATALOGING
3 S REVIS7
4 C (1 OR 2) AND 3
Comments
Since the Rules can be spelled with orwithout a hyphen, this statement retrievesboth forms
A high precision search uses nearly the same terms but scans the
identifier field for the ANCLO-AMERICAN facet.
ForT7uiation
S AACR/ID
2 S ANGLO AMERICAN CATALOGING RULES
3 S REVISION
4 C (I OR 2) AN_
Comments
After EXPAND! ng ANGLO AMERICAN as anidentifier, SELECT from the alphabeticallisting instead of typing it in yourself.
A Briefsearch might assume the following form:
Formulation #3. Briefsearch
1 S ANGLO(114)CATALOGING(C)REVISION
Step 8: Evaluating Final Results
Formulation
Number of citations round 5
Recall
Precision
98
NumberNumber o
Comments
The (C) limiter saves another step forSELECTing and then COMBINing the tworesults.
relevant citations foundcitations in answer set
Number of relevant 'tations foundfo6ndNumber o
3 (100% relevant documents3 retrieved)__
3 (60% of retrieved docu-5 monts are relevant) -
Formulations '2 and #3
Number of citations found = 3
Recall -
Precision
Number of relevant citations found 3 (100% relevant documentsNumber of citations in answer set 3 . retrieved)
Number of relevant citeumber of cit
ions-found 3 (100% of retrieved docu-ns ound 3 ments relevant)
After reviewing the recall and precision scores summarized above, it is clear
that no one formulation is favored; all perform well and are efficient and accurate.
Relevant citations:
EJ 121112 ED 108684 ED 100292
Search topic
Step 3:
TEACHER EDUCATION
111111111ftAME:AMMELAIMMAdmicAmm
,
Drug abuse including alcohol, among students of
elementary and secondary schoo.s, grades K-12. ANft1107)
Fo ulati Basic Search Lo Plannin Search S -a e
The search topic can be broken down into two facets: Drug abuse and K -1.2
students. The diagram below shows that the intersection of the two facet is
portion of the data base intended fo-'retrieval.
DRUG ABUSE
K-12 STUDENTS
Step 4: C in trie Search T
This search is simplified by the availability of ERIC descriptors for DRUG
ABUSE and for student grade level. A high recall search may contain the names
of specific drugs such as marihuana or LSD.
Step 5:
It is anticipated that the volume of literature on the topic is quite sub-
stantial; consequently, both facets will probably be introduced to the formulation.
The building block -approach is used in all three formulations below for a high
recall and high precision search, and Briefsearch; tae building block approach is
characterized by the development and union of all search facets.
Step 6: Cence- gin -arch as Irt.ut
Each search sta emen in the Briefsearch given beltow is an example of the
-g block approach as both facets, DRUG ABUSE. and K-I2 STUDENTS are present.
Search terms chosen for the Briessearch are derived from ERIC descriptors having
high postings or used frequently in descriptor phrases.
'sea
1 S ALCOHOL(C)SCHOOL(C)STUDENTS
S DRUG(W)ABUSE(C)HICHMSCHOOL
3 S DRUG(W)ABUSE(C)ELEMENTARY
4 C 1-3/OR
A high precision
Comments
The (C) limiter is used rather thanthe AND in order to reduce the numberof individual SELECT and COMBINE oper-ations
-ion relying solely on the descriptor phrase Drug
Abuse as a descriptor, identifier, or title word, to create the DRUG ABUSE t
and descriptors for grade is- shown below:
Fo lation #2
I S DRUG(W)ABUSE/DE,ID,TI
S ELEMENTARY SCHOOL STUDENTS
3 S HIGH SCHOOL STUDENTS
S SECONDARY SCHOOL STUDENTS
5 C (2 OR 3 OR 4) AND I
Comprehensive Search Saves for grade levels K-8 (Search Save #2. see appendix)
and grade levels 9-12 (Search Save 1`3, see appendix) are used in the following
high recall formulation to create the K-12 STUDENTS facet.. The DRUG ABUSE facet
is comprised of individual drug names in addition to general terms such as narcotics,
drug addiction, etc.
Formulation Comments.
I ..EXECUTE (Search Save fL_ELEMENTARY EDUCATION)
1.717641TE (Search Save forSECONDARY EDUnTI
102
Statements 1 and 2 create the setfor K -12 STUDENTS
Fo_ (cont.)
3 S ALCOHOL;
DRVO(W)ABUSE
5 S DRUG(W)AhDICTION
6 S LSD
7 S LYSERGIC
S MARI7UAN1
9 S NARCOTIC,
10 C 3-9/OR
11 C 10 AND (1 OR 2
Step F: Evaluat
Both the BriefsearchA
precision is required.
Comments
Included DESCRIPTORS and Identifiers:ALCOHOL EDUCATION, ALCOHOLIC BEVERAGES,ALCOHOLISM, Alcohol, 4 multi-term alco-hol identifiers, DRUG ABUSE, 5 multi-term drug abuse identifiers, DRUG ADDIC-TION, LYSERGIC ACID DIETHYLAMIDE, MAR-IHUANA, NARCOTICS, 2 multi-term Narcoticsidentifiers
DIALOG's truncation feature allows variantspellings of the term to be retrieved
Results
nd formulation - are favored in instances when high
Simple, efficient and requiring little typing, the
searches retrieve few non-relevant documents.
-Number of citations found = 44
Recall
Precision
Numberumber
relevant it
citations intions found 39 (60% relevant datumanswer set 65 retrieved)
Number of relevant citations faunNumber of - ions foun
Number of citations found
NumherRecall
Precision =
39
44
nts
(89% of retrieved doc=____
uments are relevant)
relevant ccit_atio5 found_ 11 (17% relevant documents1,5 retrieved)
!aimber of citations in answer set
Number of -e
Numberound 11 (100% of retrieved doe-
Search formulation #3 is prefer
connectconnect time is necessary than fo
and search Savs are processed.
ons found 11 uments are relevant)_
-- high recall searches. More online
11 or //2, as more terms are entered
all relevant documents are retrieved
with a minimum nonrelevant items.
Formulation
Number of citations found - ;5
Recall
Precision
Relevant citatio
Number of relevant citations found 65 (100% relevant documentsNumber of citations in answer set 65 retrieved)
Number of relevant citations foundNumber of citations found
65 (87% of retrieved doc-75 uments are relevant)
EJ119196 EJ118984 E1118982 EJ118979 EJ117327 EJ116766 EJ115585.EJ1155S5 EJ115581 EJ115577 EJ114039 EJ1i4035 EJ114034 EJ114033EJ114032 EJ114029 EJ110527 EJ110132 EJ108320 EJ107353 EJ107314EJ105590 EJ105584 EJ105583 ED110574 ED109914 ED109844 ED109165ED109069 ED109063 ED108016 ED107899 ED107664 ED107663 ED107618ED107572 ED107550 ED106987 ED106719 ED106706 ED104786 ED104620ED104006 ED103756 ED103746 ED103461 ED103361 ED102658 ED101866ED101499 ED101464 ED101256 ED100642 ED100610 ED100597 ED099720ED099398 ED099354 ED099338 ED098465 ED098181 ED097604 ED097234ED096577 ED095691
RURAL EDUCATION---..111MEIMMTMIMIUNN
AUL. 41ft..411.6.
Search topic #9: Vocational education of the American Indian (his-
tory, data and programs to provide this education but not
training materials to be used in these educational programs.,
(S AN=D05)
Search Objective: The requestor wants all the relevant material in the ONTAP file.
Step 3: Formulating Basic Search Lo:ic--Plannirtrateiies_
Two concepts are present in the topic: Vocational Education and American
Indians. Although the requestor is not interested in x!1111LrIg_ materials, 'this
aspect of the search will be difficult to represent in the formulation without
eliminating some relevant it__ which contain information on industrial training.
Errors in indexing the concept AMERICAN INDIANS may occur so that a set repre-
senting INDIA will be created and joined in a NOT relationship with the facet
for AMERICAN INDIANS. In this way, material dealing with American Indians but
indexed as INDIANS would be retrieved,(INDIPTS is used as an ERIC descriptor to
describe "natives of India or of the East Indies," according to the scope_note
in the ERIC Thesaurus).
The search logic is illustrated in the diagram below; the shaded area is the
portion of the data base intended for retrieval. Note that the intersection of- - _
all three sets is not wanted, rather the intersection of the two sets for VOCA-
TIONAL EDUCATION and AMERICAN INDIANS is desired in all cases where the t
INDIA does not occur,
AMERICAN INDIANS
VOCATIONAL EDUCATION
Step 4: Coffipiling Search Terms
The VOCATIONAL. EDUCATION aspect is well represented in the ERIC Th aurus
by a host of terms. Use can be made of the online thesaurus instead of typing
in a number of terms and risking typing For a high recall search,
several expressions must be entered to represent AMERICAN INDIANS, particularly
individual tribal names Formulation #2 below employs Search Save 115 (see appen-
dix) on general Indian names; however, a more comprehensive Search Save contain-
ing nearly all tribal names of North American Indians could be devised for high
recall searches.
Step 5: Ordering
e search objective, to retrieve all relevant citations in ONTAP, plays a
part in determining what approach to search strategy should be employed. Since
neither facet, American Indians nor Vocational Education, could represent the
topic alone without containing much non-relevant material, it is expected that
both facets will be involved in the formulation. Thus, the building block
approach appears to be a suitable choice; however, the successive fractions
approach could also be employed to execute a comprehensive subsearch for American,
Indians; in this way, items indexed erroneously and/or dealing with American
Indians but not referring to them specifically are retrieved.
Step 6: Conceftualizin to the Retrieval
Formulation #1 is an example of the successive fractions approach. In the
first four search statements, the facet AMERICAN INDIANS is built by using Search
Save #5 and the search terms India and Indian; then, the set for INDIA is parti-
tioned from the AMERICAN INDIAN set. The following diagram illustrates this
process; the shaded area represents that portion of the data base which is treated
in subsequent statements of the formulation; the unshaded area is, literally,
discarded.
containing:
Search Save results. Indian results
Set containing:
3. India
The full formulation is given below:
Formulation # Comments
1 .EXECCIE_ (Search Savefor AMERICAN 1. LANS)
2 S INDIANStatements 2-4 draw out materialon INDIAN and INDIA. The set for3 S INDIAINDIA is then COMBINEd in a NOT
4 C (1 OR NOT 3 relationship with the AMERICANINDIAN facet in order to eliminateitems on India which may be non-relevant
5 S INDUSTRIAL(W)ARTSStatements 5-9 comprise the facetof VOCATIONAL EDUCATION
6 S INDUSTRIAL(W)XDUCATION
7 S VOCATION?
8 S INDUSTRIAL(W)TRAINING
9 C 5-8/OR
10 C 4 AND 9
The Briefsearch formulation, a neat and efficient single search statement,
retrieves all but one relevant item.
Briefsearch
1 S AMERICAN(F)INDIANS(C)VOCATIONAL
Comments
DIALOG processes word proximity state-ments which range in length from a fewcharacters to as many as Fit on a singletyped line
101
Step 2: Evaluating Final Results
Before reviewi54 the evaluation of the riefseaBriefsearch and the high recall
formulation, how well does your search perform In terms of recall and precision?
Twenty-one citations are judged as being relevant. Did you retrieve all the
relevant documents?
The results of the high recall formulation reveal that all relevant citations
are retrieved.
Formula ion #1
Number of citations found g, 35
Number oRecall =
Number of citations in answer setlevant citations mind
PrecisionNumber of levant citations found
Number of citations found
21 (100% relevant documents21 retrieved)
21 (60% of retrieved doc-35 uments are relevant)
On the other hand, the Briefsearch failed to recall one relevant citation.
An error in indexing accounts for this, as the descriptor Indians which refers
to the "Natives of India or of the East Indies" according to its scope note, is
assigned to citation ED 097483 dealing with Indians in Arizona. In order to re-
trieve this item, it is necessary to execute the NOT operation as described in
formulation #1 cor enter individual tribal names of North American Indians, Re-
call and precision for the Briefsearch are given below:
Briefsearch
Number of citations found 25
Number of relevant citations found 20 (95% relevant documentsRecallNumber of Jtations anAwer set 21 retrieved)
Number of relevant citations found 20 (80% of retrieved doe -Number of _ ons found 25 uments are relevant)
The question of which formulation is preferred in this case is left to you,
the searcher.
108
Relevant citations:
EJ 113861 EJ 112180 ED 1C9445 ED 108822 ED 108816 ED 108797 ED 108464ED 107424 ED 107396 ED 107395 ED 101313 ED 101167 ED 100542 ED 100541ED 100539 ED 100437 ED 098000 ED 097483 ED 097163 ED 096073 ED 095410
109
EARLY CHILDHOOD EDUCATION
SCIENCE, MATHEMATICS, ANDENVIRONMENTAL EDUCATION
Search topic #10: ,lean ?Jager s theories, and thought processes
or language development of children. (S AN=M03)
Search Object e: All relevant citations are wanted by the re-
questor; no journal articles are desired.
Step 3: Formulating Basic Search Plannin Search Str e _e
Piaget represents one facet of the search. Piagetian theories, language
development and thought processes comprise aspects of the second facet (this
second facet is referred to as THEORIES in this discussion). 'Children forms
the third facet as the requestor desires information on theories of Children,
thought processes of Children, or language development of Children.
Step 4: Compiling the Search Terms
To incorporate Piagetian theories in the formulation, the searcher must be
aware of the vocabulary and terminology associated with Piaget's work. The need
to include many variant or synonymous expressions for the theories adds to the
compleXity of this search. The ERIC Thesaurus and Thtirusof:aoical
Terms are valuable sources for gathering such search terms. In search formula--
tion 114, terminology for thought processes, language development and Piagetian
theories make up a substantial portion of the search.
Step 5: OrderingQt4put,
Since the output is limited to publications available from ERRS, the second
step in every formulation is the LIMIT command; this restricts the results of the
first facet to the appropriate body of documents and allows the searcher to monitor
preliminary results in the light of one of the search objectives.
111
Beginning with the most specific aspect of the search, PIAGET, the
searcher may decide to terminate the search without adding another facet to
the formulation. Examples of the most specific aspect first, building block
and Briefsearch are given. The COMMENTS section of each formulation explains`
the searcher's decisions during the preliminary evaluation process.
Step 6 : Concert the Searchas In utto the Retrieval System
A Briefsearch, pertorm0 initially before a complex formulation, gives a
rough estimation oaf the si of the final output. Records from Briefsearch
results can be reviewed in order to collect additional searching vocabulary.
Briefsearch
1 S PIAGET(C)COGNITIVE
2 S PIAGET(C)CHILDREN
3 C 1 OR 2
4 LIMIT 3/AVAIL
At this point 72 citations are retrieved. Searching on the most specific
facet, PIAGET, and comparing the results of this statement with that of the
Briefsearch gives us a general idea how large the final set may be.
ormulation 112
1 S PIAGET?
2 LIMIT 1/AVAIL
Comments
Included DESCRIPTORS and identifiers: Piaget,17 multi-term Paget identifiers, Piagetian, 9multi-term Piagetian identifiers.
This search retrieves 93 citations. Since therequester wants all relevant items; the searchermay consider terminating the search at thispoint; having determined that the volume of out-put is small enough for manual review.
The addition of the facet CHILDREN to the formulation reduces the output to
77 citations. This size of output may be considered sufficient to forego the
introduction of another facet to the formulation.
Formulation #3
3 . EXECUTE (Search Savefor CHILDREN)
C 2 AND 3
The fourth formulation is com
It requires much typing so that using
of employing free text searching.
Formulation #4
1 S PIAGFT?
2 LIMIT 1/AVAIL
3 S AFFECTIVE(W)BEHAVIOR
4 S COGNITIVE(W)ABILITY
5 S COGNITIVE(W)ABILITIES
6 S COGNITIVE(W)DEVELOPMENT
7 S COGNITIVE(W)DEVELOPMENTS
8 S COGNITIVE(W)OBJECTIVE
9 S COGNITIVE(W)OBJECTIVES
10 S INTELLECTUAL(W)DEVELOPMENT
11 S INTELLECTUAL(W)DEVELOPMENTS
12 S MENTALMDEVELOPMENT
13 S FERCEPTUALOODEVELOPMENT
14 S ECHOLALIA
15 S FLES
16 S READING(W)HABIT
17 S READING(W)HABITS
18 S HUMAN(W)DEVELOPMENT
19 S LANGUAGE?
20 S NUCLEATION,
21 S PSYCHOLINGUISTIC?
22 S READING(W)DEVELOPMENT
23 S_LINGUISTIC?
Search Save 174 (see appendix) is usedt:.r the CHILDREN facet
ed of two facets, THEORIES and PIAGET.
the online thesaurus may be favored instead
Commen
Just like in formulation 113, the first twostatements are taken from formulation #2and make up the entire facet PIAGET.
Included DESCRIPTORS/Identifiers:AFFECTIVE BEHAVIOR
COGNITIVE ABILITY
Cognitive Abilities Test
COGNITIVE DEVELOPMENT/Cognitive Develop-ment Research Tools
COGNITIVE OBJECTIVES
INTELLECTUAL DEVELOPMENT
MENTAL DEVELOPMENT
PERCEPTUAL DEVELOPMENT
ECHOLALIA
FLES, FLES GUIDES, FLES MATERIALS, FLESPROGRAMS, FLES TEACHERS/FLES Objectives,FLES Program
READING HABITS
HUMAN DEVELOPMENT/3 multi-term HumanDevelopment Identifiers
LANGUAGE, over 60 multi-term LANGUAGEDESCRIPTORS, 31 LANGUAGES DESCRIPTORS/50 multi-term Languav Identifiers
NUCLEATION (LANGUAGE LEARNING)
PSYCHOLINGUISTICS
READING DEVELOPMENT
LINGUISTICS, 13 multi-term LINGUISTICS andLINGUISTIC DESCRIPTORS/26 multi-termLinguistic and Linguistics Identifiers
Formulation #4 (Cont.) Comment
24 S SPEECH(W)DEVELOPMENT RETARDED SPEECH DEVELOPMENT
25 S SPEECH(W)HABITS SPEECH HABITS
26 S VERBAL(W)DEVELOPMENT VERBAL DEVELOPMENT
27 S VOCABULARY(W)DEVELOPMENT VOCABULARY DEVELOPMENT
28 S HANDWRITING(W)DEVELOPMENT HANDWRITING DEVELOPMENT
29 C 328/OR Cumulative set for cognitive andlanguage development
C 29 AND 2
Imposing the '!acet PIAGETIAN THEORIES reduces the final set to 71 citations.
At this point, the searcher terminates the search in order to avoid losing rele-
vent material. If a decision is made to include the CHILDREN facet, use could
be made of the Search Save demonstrated in formulation #3.
30 .EXECUTE (Search Savefor CHILDREN)
31 C 30 AND 29
After the CHILDREN facet is incorporated, the total number of retrieved
citations is 60.
(Reader: In light of the requestor's desire to retrieve allrelevant citations 'regardless of cost, which strategy do youprefer?
Step 8: Evaluating Final Results
Formulation #2 in which the most specific facet of the topic is searched
retrieves all 81 relevant citations. It is the only formulation of those given
which achieves 100% recall.
Formulation
Number of citations found a 93
Recall
Precision
Number oNumber of citations I
eleven on ound 81 (100% relevant documr nanswer set
Number of relevant citeNumber of cltati
ons foundons found
81 retrieved)
81 (87% of retrieved doc-93 uments are relevant
all
The Brief search achieves high precision and high recall but does not retrieve
s relevant to the topic. But, used as a brief surrey of the file, it
aids in determining whether to impose a second facet to the formulation.
Brief search
Number of citations found = 72
Number nf relevant cNumber of citations
Recall bons und 66 (81% relevant documentsanswer set 81 retrieved)
Precision Number of relevant citations found 66 (92% of retrieved doc-=Number of citations found 72 uments are relevant)
Two facets, CHILDREN and FIAGET, of the topic make up formulation #3.
Imposing the CHILDREN aspect into the formulation eliminates 14 relevant doc=
uments from the output.
Formulation #3
Number of citations found = 77_
Number of relevant citations found 67 (83% relevant documentsRecall = __-- i
relevantNumber of,citatio-:s in answer set 81 retrieved)
Precision Number of relevant citations found 677_ (87% of retrieved doe-Number of-citations found 77 uments are relevant)
All facets of the topic are included in the building block approach,
formulation 114. If the search is terminated before the CHILDREN facet is
included in the formulation, 71 citations are retrieved; after the imposition
of the CHILDREN facet, 60 citations are retrieved.
:Formulation #4. ("Before" CHILDREN facet, statements 1-31)
Number of citations found = 7i
Number of relevant citations found 71
Number of citations in answer set 81 retrieved)Recall
Precision
% relevant documents
Number of relevant citationNumber of citations
foundnand
7171
(100% of retrieved doesuments are relevant)
115
Formulation 44. ("After" CHILDREN facet, statements 1-33)
Number of citations found m 60
Recall = Number of relevant _citeNumber o
Precision =
dons in answeound 60 (75% relevant documenset 81 retrieved)
umber of evant citations found 60 (100% of retrieved doc-Number of citations found 60 uments are relevant)
As more facets are added to the formulation in the third and fourth
examples, therecall ratio decreases. Since high recall is one of the search
objectives, formulation 112 is preferred. Not only does it achieve a high
recall score, but it is efficient and rapid in terms of online connect time
and pre - search preparation.
Relevant citations:
ED 110469 ED 110190 ED 110186 ED 110173 ED 110159 ED 109196 ED 109149ED 108957 ED 108893 ED 108864 ED 108862 ED 108751 ED 108748 ED 108745ED 108404 ED 107562 ED 107385 ED 107366 ED 106721 ED 106147 ED 106131ED 106122 ED 106117 ED 106042 ED 105979 ED 105976 ED 104739 ED 104552ED 104530 ED 104516 ED 104051 ED 103296 ED 103133 ED 103129 ED 103127ED 103106 ED 103100 ED 103095 ED 102240 ED 102111 ED 101849 ED 101847ED 101837 ED 101565 ED 101488 ED 100704 ED 100536 ED 100529 ED 100528ED 100521 ED 100510 ED 100488 ED 100485 ED 099811 ED 099767 ED 099415ED 099196 ED 099122 ED 099112 ED 099098 ED 098793 ED 098517 ED 098490ED 098082 ED 097991 ED 097984 ED 097977 ED 097970 ED 097959 ED 097957ED 097379 ED 097261' ED 097129 ED 097110 ED 097109 ED 097102 ED 096016ED 096000 ED 095999 ED 095997 ED 095996
116
HANDICAPPED ANDGIFTED CHILDREN
MEIMMINIKNOMMISIMII
Search topic #11: Identification of the Gifted.
Search objective: The requester needs immediate results
_
Step 3: FormL11.!Ilgg Basic Search 1,o,ic Plannint Search Strate
The search topic is broken down into two facets: identification and gifted.
The diagram below illustrates search logic:
GIFTED IT BNTIFICATION
Step 4: Compiling the Search Terms
Search terms can be gathered from the ERIC Thesaurus, Ro et's Thesaurus,
and the Thesaurus of ps.x212212aLe41 Terms. However, the requestor's desire for
immediate results curtails lengthy pre-search preparation on the part,of the
searcher. The search topic is comprised of two heavily posted descriptor.
these terms also represent both search facets. Thus, the searcher can use the
terminology of the search topic as search terms rather than spend time examining
various thesauri.
Step 5: Ordering Output
The searcher, obliged to produce immediate results must curtail pre-search
preparation; selecting the citation pearl growing approach to search strategy
allows the searcher to review preliminary results of the search in order to
gather additional search terms so that preparation before the online interaction
is not essential. Moreover, the searcher can revise search strategy when a
search term retrieves non-relevant material.
High precision is an unstated objective in this search as the requester's
demand for i ediate results requires online printing of the output. Online
printing of a large percentage of non-relevant Its is costly in terms of
online connect time and requires that the information seeker spend time wading
through non-relevant material rather than concentrating on the relevant items.
In addition, online printing of the format that provides only the information
essential to the requester is suggested.
Step 6: Conce ualien the Search _as t to the Retrieval_System
Using the citation pearl growing approach, the searcher begins with a
very direct search on the most spedific search terms. Expressing the tooic
in the terms used by the requester, identification and gifted, is limited to
retrieving citations whose titles contain these terms. The searcher then prints
a few citations in format 02, checking the descriptor field in order to gather
other search terms.
Formulation 01
S GIFTED(F)IDENTIFICATION/TI
TYRE 1/2/1-2
Comments
Five citations are retrieved; the searcher,confident that all five are relevant,checks the index terms in order to addmore search terms to the formulation
EJ103421 UD503083THE IDENTIFICATION OF MENTALLY GIFTED, "DISADVANTAGED" STUDENTS AT
THE EIGHTH GPADE LEVELFITZGIBBON CAROL T.JOURNAL OF NEGPO Etfl 43 1 53-66: WDESCRIPTORS: +JUNIOR HIGH SCHOOL STUDENTS/ +TALENT IDENTIFICATION./J--
+DISADVANTAGED YOUTH/ +URBAN SCHOOLS/ NEGRO STUDENTS/ GIFTED/ TALENTED..STUDENTS/ ECONOMICALLY DISADYANTOGED/ PREDICTIVE MEASUREMENTIDENTIFIERS: +CALIFORNIA
ED1,08279 - EA007191GIFTED STUDENTS: IDENTIFICATION TECHNIQUES ANDDOOB, HEATHER S..'EDUCATIONAL RESEARCH SERVICE, WASHINGTON, D.C.ARP 75 61P.AVAILABLE FROM EDUCATIONAL RESEARCH SERVICE, INC. 1815 NORTH FORTMYEP DRIVE, ARLINGTON' VIRGINIA 22209 ($%.00, PAYMENT MUST ACCOMPANY
ORDERS $10.00 OP LESS)DOCUMENT NOT AVAILABLE FROM EMS
PROGRAM O GANIZf ID
DESCRIPTOR S:.. ,ABILITY :IDENTIFICATION/ -BIELIOGRAPHIES,....ELEMENTARY&.014DARY,E0061TION2 +GIFTED/- +PROGRAM CONTENT/ -*PROGRAM-DESCRIPTIONS/+SUPERIOR STUDENTV :.TALENTED STUDENTS/-7ALENTIDENTIFICATION-
term "Talent" from the debcriptor field, the se
the strategy in the second formulation restricting all search term
in desid-ortifier-mr-tirIP_UR_Lia,
Formulation #2
S- GIFTED/DE ID,TI
2 S TALENT/DE,ID,TI
S TALENTED /DE,ID,TI
4 S IDENTIFICATION /DE,ID,TI
5 C (1 OR 2 OR 3) AND 4
cher revises
to occurrences
Employing the same search terms. as formulation 112, the following high recall
_for ulationsearches all subject Conveying fields rather than being limited to
the title and assigned index term fields.
ForMulation 113
I S GIFTED?
2 S TALENT?'
3 S'IDENTIE?
4 C (1 OR 2) AND 3
Comments
Included DESCRIPTORS: GIFTED, IDENTIFICATION;ABILITY IDENTIFICATION, 4 Multi-term_IDENTI-FICATION DESCRIPTORS, TALENT IDENTIFICATION,TALENT,'3 multi-term TALENT DESCRIPTORS,TALENTED STUDENTS
The Brie search,- a single search statement, resembles the -lording of the
search topia.,
ief &arch
S IDENTIEICATION(F)GIFTED
-Meader: - Keeping in mind that high precision is a searchobjectiveiaawell as delivning output to the requester as-soon as possible, which formulation above fulfills bothobjectives?
Evaluating- Final Results
The answer set for the search topic is not available in the ONTAP file.
The results of your search can be compared with the list of ERIC documents
cited by accession number) concluding this discussion.
The performance of the first formulation not evaluated. Intended-'as a
demonstration of how thecitation pearl growing approach can-be employed to
gather search terms and relevant items, the formulation is actually integrated
into the second example.
Formulation #2 achieves high recall; as intended, but fails to retrieve
all- relevant items.
Number of citations found =1,8°
Recall Number of relevant cit_atlons found 16 (61% relevant documents=Number of citations in nswer;set 26 retrieved)
Number of levant citations found 16 (89% of retrieved_doe-18, uments -are relevant)
Formulation #3
Number of citations
an example of a high recall formulation, retrieves all
relevant items; however, the output contains many non-relevant items, so that
this formulation is not the preferred mode for representing the topl.c.
Formulation #3
Number of citations found 45
Number of relevant citations f-und 26 (100% relevant documentsRecall =
Precision
Number of citations in answer set 26 retrieved)
Number of relevant_citations found 26 (58' of retrieved doe-/6 uments are relevant)Number of citations found
The evaluation of the Brief search follows. Although it retrieves a larger
number of relevant items than the second formulation, eight non-releVant citations
1201 J_ 0
are included in the output sc, that formulation 2 'is favo for representing
the topic in light of the search objectives,
Brief search
Number of citations found 30
, Recall
precision
Number of relevant citations found 22 (85% relevant documerNumber of citations in answer setP
Number of relevant_citarions foundNumber of citations found
Relevant citations:
EJ 110029 EJ 109349 EJ 104281ED 109863 ED 108278 ED 107746,ED 104971 ED 104965 ED-104095ED 102773 ED 102772 ED 102771ED 097788 ED 096784
26 retrieved)
(73% of retrieved documents are relevant)
EJ 104226 EJ 103421 ED 109867'ED 106140 ED 105687 ED 105329ED 104094 ED 104039 ED 102795ED. 102765 FD 100102 ED 100072
HIGHER, EDUCATION
Search topic #12: Higher education enrollment projections for
the near future (1975-1995).
.Search Obj ctive: To retrieve all journal literature on this topic.
=Step 3: Formulat c Search Lo ic Plannin Search S at ies
There are three concepts present in the search topic: higher educntioni
ollmenCand projections. Search logic is illustrated below:
Higher Education Enrollment
Projections
Step 4: Com ilin the Search Terms
:The ERIC Thesaurus is suitable for compiling search terms for the facets
HIGHER' EDUCATION and ENROLLMENT. Termg synonymous with PROJECTIONS can be
gathered from Ro et's'Thesaurus as well as a dictionary since the descriptoro
Project, is associated with /UMW-Visual equipment,
Step 5: Ordering Output
The requester only desires journal literature so the searcher's first task
during the online interaction is to apply the LINITALL feature'. Thus, confronted
with decisions when online, the searcher can assess preliminary results more
Accurately. If the LIMIT capability is aPpended_to the formulation as a final
step, the size or contents of the final set may not be satisfactory, causing
the searcher to "backtrack" or reformulate the search._
123
High recall and Briefsearch formulations are covered in Step 6. The
building block approach is the favored mOde for repre5enting search strategy
as all three faCeta have heavily posted search terms. Restricting the formula-
lion to two facets results in retrieving a high volume of non-relevant material.
Step 6: Conceptualizing the Search a nu o the Retrieval S
In order to compile additional search terms or ?o perform a survey of the
file, a Briefsearch can be formulated.
Briefsearch
LIMITALL/EJ
1 S HIGHERMEDUCATION(F)ENRO
2 S PROJECTION?
3 Cl AND 2
Comments
282 citations are retrieved as aresult of the first search state-ment making it necessary to imposethe third facet into the formula-tion.
The high recall formulation wnploys Search SaveV10(see appendix) to
represent the HIGHER EDUCATION facet. As an example of the building block
approaCh, the following formulation is not terminated until all facets are
ratroduced into the strategy.
?ormulation
LIMITALL/EJ
1 S .EXECUTE II
2 S ENROLLMENT?
3,S PREDICT?
4 S PROJECTION?
S FORECAST?
6,S FUTUR?
- 7 C 3 =6 /OR
8C 1 AND 2 AND 7
(Search Save forHIGHER EDUCATION)
Comments
The decision to impose the thirdfacet is made after COMBINingstatements 1 ANA' 2, and re-trieving over 500 citations.
Step 8 Evaluating Final Results
The answer set to this topic is not stored in the ERIC ONTAP file, however,
you can compare your search results with the list of relevant citations concluding
this discussion. A citation is determined relevant only if it makes enrollment
_ions, and not if it treats- enrollment projections from a historical per-
spective or is about making enrollment projections.
Formplation #2 achieves 100% recall, retrieving all 22 relevant documents in
the ERIC ONTAP file.
pre
Formulation 2
'-Number of-. citations tound m 27
Recall
cision
umber of relevant citations found 22 relevant documentsNumber of citations in answer set 22, retrieved)
Number` relevantNumber o citations found
as found _22_ (81% of retrieved doe-27 uments are relevant)
The Briefsearch retrieves 17 items, all of which are relevant. It would be
:red in cases when high precision and/or low cost search objectives.0
liTiefsearch
Number of _citations found = 17
Recall = Number of relevant ation found 17 (77% relevant documents,Number of citations in answer 22 retrieved)
Precision
Relevant
Number of levant cNumber of chat
ound 17 (100% of retrieved (inc17 uments are relevant)
EJ 117830' EJ 114796 EJ 114396 EJ 114216 EJ 114215 EJ 114214
ET 112284 EJ 109179 EJ 108126 EJ 108036 EJ 107561 EJ 107550
EJ 107306 EJ 106061 EJ 105868 EJ 104496 EJ 104491. EJ 104490
EJ 104441 'EJ 1,02836, EJ 102829 EJ 102812
JUNIOR COLLEGES
Search topic #13: Junior, and community college enrollment pro--
jections for the near futUre (1975-1995).
Search Ob e ve: Th information seeker wants the output restricted to ERIC
report literature so that all journal articles are excluded.
Step 3: PormulnBasic Search Logic -s Planning search Strategies
This topic, like the one preceding on higher education Rnrollment projections,
is comprised of three facets: Junior Colleges, Enrollment and Projections.
Step 4: Compiling the-Seardh Terms
The ERIC Thesaurus, Roger's ThesaUrus, and a dictionary are suitable sour
for gathering search terms. Instead of formulating g subsearch for the JUNIOR
COLLEGES facet, the ERIC Clearinghouse on Junior Colleges could be substitUtedi
as that_clearingheuse processes material dealing specifically with Junior Colleges.
Step 5: Ordering Output
Only ERIC report literature is desired so the LIMITALL command is applied
as one of the first activities duringthe online Interaction. The building block .
and successive fractions approaches discussed below, are examples of high recall
and high precision formulations, pectively.
Step 6: Con e ualizin the Search o the Retrieval S etetn
In the successive fractions approach, the LIMITALL, applied as the first
e0 in the search, takes the 'initial !'bite" of the file. This set Is then"
strieted to documents,indexed by the Junior College Clearinghouse/ The remaining
facets are incorporated into the strategy and the search terminated when the
,final set is satisfactory to the searcher.
Formulation #1
LIMITALL/ED
1 S CH=JC
2 S ENROLLMENT?
3 C 1 AND 2
4 S PROJECTION?
5 C 3 AND 4
Coruments
All ERIC reports indewl by the JuniorColleges Clearinghouse form afacet for, JUNIOR COLLEGES.
At this point 140 citations'are retrievedand the Searcher continues, adding thethird facet to the search.
The -search is terminated after the third -
facet is imposed; the final set con-tains 21 citations,
Formulation #2 employs Search Save #9 to build the JUNIOR COLLEGES-facet..
As a high recall formulation, it is an example of the building block approach.
-rmulat_on #2 Comments
LIMITALL/EJ
1 .EXECUTE Search Save for
'2 S ENROLLMENT?
3 S PROJECT?
4 S FUTUR?
5 S ESTIMAT?
6 S FORECAST?
7 C 3-6/OR
8 C 1 AND 2 AND 7
Step 8: Evaluatin
JUNIOR COLLEGES)
Final Results
Searching is not Confined to thedescriptor field but encompassesall subject-conveying fields ofthe ERIC Record.
The answer set o the topic is not available in the ONTAP file; the list of
relevantr-cititions folio _ thisdistussion. You can figure redall and precision
percentages using ',the formulas summarized below in the evaluation of formulations
#1 and #2.
if high recall is a search objective, the second formulation is preferred.
Even though all 26 relevant citations ire retrieved, the ratio of non-relevant
.citations 1.6theoutpUt to relevant citations is nearly 2 to 1. In order to.
capture all relevant citations in the ONTAP file, as many search terms as possible
are used to represent the facets JUNIOR COLLEGES and PROJECTIONS. This,
addition to free tent searching, may account for the low precision of this
formulation.
In contrast, formulat on,#1, almost a-Briefsearch considering its ease
and efficiency, achieves high precision and would be favored in cases requiring
a high precision, low cost formulation.
Number of citations found = 21
Recall Number of relevant citations found 16 (62% relevant documents=
Precision =
Number of
Formulation 2
ions in answer set 26 retrieved)
NUmbe- of relevant citations found 16 (76% of retrieved doe-Number of citations ound ' 21 uments are relevant)
Number of citations found = 68
Number of relevant citations found 26 (100% relevant documentsNumber of citations in answer set 26 retrieved)
Number of e event citations found 26 (38% of retrieved doc-_ati©ns. foundfound 68 uments are relevant)
Recall
Precision
Relevant citations.:
'ED 110153 ED 110152 ED 110122
ED 105958 -ED 105919 .ED 105918
ED 101792 ED 101786 ED 100431
ED 099062 098887 ED 098881
ED 095948 ED 095782
ED 110145
ED 105913
ED 100250
ED 097035,
ED'110107 ED 109582
ED 104213 ED 103065
ED 099082 ED 099080
EDr096895 ED 095971
URBAN EDUCATION1.11111MMAIMMIENFMININOU
Search topic 14: The Family background of Puerto Rican children
Step 3: Formulating Basic_ Seasstj.11EUL=Planninz Search Strate
Three concepts are present in the search topic: 1:(Puerto Ricans, children
and fdmilies. The diagram below illustrates the search logic:
Step 4: Compiling the Search Terms
Searching vocabulary can be compiled from the ERIC Thesaurus Below are
listed the-three facets and terms'which are suggested for inclusion in a for-
mulation.
PUERTO RICANS
Puerto Rican -sPuerto Rico
fi
Family -iesHousehold -sMother.-aFather -s,Parent -s
CHILDREN
Child -ten .
Boy -sGirl -sJuvenile -sTeenage -r -rsYouth -sBaby -lesInfantAdolescent -ce -sToddler -s
Step 5: Ordering Output
The suggested method o aPproaching this topic is beginning with,the Most
specificfacet, PUERTO RICANS, which is a named object and easy to represent,
and adding the Other facets to the formulation until the results are satisfactory
to the searcher. This mode of search strategy is actually comprised of two
app aches, most specific facet first and successive fractions. The building
plock_approach could al.:o be employed as an approach to search strategy; both
approaches are illustrated below
Step 6: Conce ual zin In ut to the Re _eval S s
Formulation /11 is a combination of the- most specific facet first and successive0
fractions approaches. The subsearch for PUERTO RICANS is executed first so that
the searcher can find out how many items are in the file containing variant f
of Puerto Ricans before formulating. another subsearch., Since the set for PUERTO
RICANS is considered-too large, the searcher continues, evaluating preliminary
results after incorporating additional facets o the formulation.
Formulation
1 SPUERTO(W)RICO
2 S PUERTO(W)RICAN
3 S PUERTO(W)RICANS
4 C 1-3/OR
'5 .EXECUTE
6 C 4 AND 5,
7 S FAMILY
8 S FAMILIES
S HOUSEHOLD?
10 S PARENT?
11 S MOTHER?
12 S FATHER?
13 C'7-13/OR
-14 C 6 AND 13
(Search Save forfor CHILDREN)
ants
161 citations are retrieved at this point,and the searcher continues rather thanterminating the search.
The search could be ended here; 54 citationswould comprise the output. However, thesearcher imposes the third andiast facetof the topic, FAMILIES.
The output contains 23 citations.
9
The Brie =ch contains the terminology used in the search topic:
Puerto Ricans, families and children. Since all three facets are employed,
it represents the building block appro
Brief search
1 S PUERTO(W)RICANS(C)CHILDREN
2 S FAMIth
3 C 1 AND 2
Step 8: Evaluatin Final ,Results
Formulation a achieves high recall, retrieving all seven relevant it
from the ONTAP file. However, since the information seeker must ade through
much non-relevant material, the Briefsearch is favored when high precision is
a search objective.
ulation #1
Number of citations found = 23'
Number of relevant citations found_Recall =.Number of citations in answer set
Precision
arch
(100% relevant documentsretrieved)
Number of relevant citations found 7 (30% of :retrieved doc-23 uments are relevantNumber of citations found
Number of citations found = 5
Recall =
Precision
nations found_ 3 (43% relevant documentsNumber of citations in answer set 7 retrieved)
Number of relevant c
Number of elevant citations found 3 (60% of retrieved do c65 uments are relevant)Number of citations found
Relevant citations:
ED 107742 ED 106383 ED 105055 ED 103739 ED 103381 ED 103371
ED 099276
TESTS1 MEASUREMENT,AND EVALUATION
Search topic 5: Screening preschool children for potential
learning disabilities or mental handicaps. Items,discusa
screening for physical handicaps are net wanted.
Search Objective: Only journal articles are wanted; also, the info_
seekerseeker` desires high precision and a low cost sea
p 3: nulatiorrSearch Strategiea
There are three concepts in the search topic: screening, preschool and
handicaps. The portion of the data base intended for retrieval must contain all
three facets, so thai the three facets are joined by the,AND operator in the
search logic which is illustrated below.
Handicaps Screening
Since information on physical handi6aps is not wanted, the search logic
may be represented in another way. Inthis case, PHYSICAL HANDICAPS forms a
Fourth facet. The blackened portion in the diagram below contains the facet
PRESCHOOL, SCREENING and PHYSICAL HANDICAPS. Since this combination is. NOT
desired, it is partitioned from the retrieved set using the NOT operator.
Rather, the shaded area containing PRESCHOOL, SCREENING and MENTAL HANDICAPS,
NOT PHYSICAL HANDICAPS is, desired.
Physical Handidaps
Mental Handicaps
Step 4: Compiling the Search Terms.
The ERIC Thesaurus is a rich source for- search terms representing all three':
facets. Since two search facets make use of the Search Saves in the appendix
it is evident that -this topic is made up of concept groups which
in ERIC searches. Check the Search Saves for PRESCHOOL and MENTALLY HANDICAPPED-_
as you compile search terms your.own formulation,in order to gather
vocabulary,
Step 5: Ordering Output
Sinc,e all three facets are heavily posted, it is expe ed that all facetS,
will be,ihtroduced into the formulation. The building block approach may bez-_,
suitablOuode for formulating the search, as subsearches for
developed and combined into the formulation. Since low cost is a search objec=_T-,,
every facet are
tive, the building block approach may not be A wise choice as the combination'',
of all facets may result in no or so few postings that the searcher may have to
"backtrack" Or reformulate the,search; consequently, the searcher may dedide to
begin the,--iline interaction with a Briefsearch. Employing all facets in the
Briefsearch, the searched may then elect, depending upon, the results o follow
uplwith a building block appro'ach. The successive fractions Approach is an
alternative if the Briefsearch results in feW postings.
Another means of educing online connect time is to apply the LIMITALL
command as the first statement in the formulation; in this way, one of the
output specifications, i.e., restricting the output _to journal articles, is
satisfied and the remainder of the search formulation is performed in the
appropriate subfile, reducing pocessing time.
High precision is also a search objective. Adding the fourth facet,
PHYSICAL HANDICAPS, the formulation as a concept not desired may reduce
the volume of non-relevant items _ the output. Favoring controlled term
and title searching instead of free text may also result in fewer non-relevant,
_ems in the retrieved set.
is important that you, the searcher, keep in mind the search objectives
stated by the information seeker when selecting an approach to search strategy.
In Step 6, four formulations are covered; when you have reviewed each one,.
select the one you feel satisfies the requestbr's needs or develop your own
formulation to achieve the search objectives.
Step 6: Conceptualizin in the an-ua e he Retrieval S-s
The Briefsearch'is used initially in order to determine whether to develop
more complex formulations in the building block or successive fractions approach.
All three facets are present in the following formulation:
Brief search
LIMITALL/EJ
1 S EARIY(WHILDHO D(C)TESTS
2 S HANDICAP?
3 C 1 AND 2
Eighteen citations are retrieved by the Briefsearch. The searcher engages
the building block approach in the next example. High recall is a primaty goal
in this case, as two Search Saves are employed-to represent the HANDICAPS and
137
and PRESCHOOL facets and free text searching Is used throughout the formulation.
ula- ion #2
LIMITALL/EJ
1 .EXECUTE
2 .EXECUTE
3 S TEST? ?
4 S TESTED
5 S TESTING
6 S IDENTIF?
7 S SCREEN?
8 S DIAGNOS?
9 C 3-8/OR
10 C 1 AND 2 AND 9
11 S PHYSICAL?
12 C 2 NOT 11
13 C 12 AND 2 AND 9
Comments
(Search Save forPRESCHOOL)(Search Save tarMENTALLY HANDICAPPED)
Statements 3-8 fort the facet for SCREENING
At this point; 62 citations are retrieved;-it is decided to incorporate the fourthfacet, PHYSICAL HANDICAPS, to the for-mulatiou in order to exclude items inthe output which treat physical handi-caps, a topic not wanted.
In statement 12, items containing the rootPhysical- are singled out and discardedfrom the set On MENTAL HANDICAPS.
:fifty -nine citations are retrieved and thesearcher has decided to end- the formu-lation at this point.
Instead of sing the comprehensive Search Saves, both formulations 1/3 and
#4 incorporate the terminology used by the information seeker to express the
search topic. Only the search terms, Test, its variant forms, and Early Child-
, hood are added to the searching vocabulary in order to increase recall. In
formulation #3, free text searching is used.
Formulation
LIMITALL/EJ
1 S EARLY(W)CHILDHOOD
2 S PRESCHOOL?
© 3 S LEARNING(W)DISABILITIES
1 38
Formulation (con'i
4 S HANDICAP?
5 S TEST?
6 S SCREEN?
7 C (1 OR 2
AND (5 OR 6)
(3 OR 4)
Since truncation is used freely with such"common" roots (as opposed to unique)as Test- and screen-, it can be expectedthat false drops will occur and affectprecision adversely.
Thirty citations are retrieved in formulation-113. Using the same search
terms, formulation #4 is limited to searching the controlled term fields and the
title. It is anticipated that few non-relevant items will'be retrieved by this
formulation.
Formulation (4 Comments
LIMITALL/EJ
1 S EARLY(W)CHILDHOOD/TI,ID,DE
2 S PRESCHOOL/TI,ID,DE
3 S LEARNING(W)DISABILITIES/TI ID,DE
4 S HODICAPPED/TI,ID,DE
5 S TESTS/TI,ID,DE..
6 S SCREENING/TI,ID,DE
7 C (i OR 2) AND (3 OR 4) AND OR 6) Twenty-two citations a- retrievedat this point.
(Reader: Remembering the search objectives, high precision andlow cost, which formulation of those given above would you chooseas satisfying these objectives? Have you developed a betterstrategy to meet the objectives?
Step 8: Evaluatinalinal Results
Twenty-eight ERIC documents are judged as being relevant to the search
topic; the list of these 28 citations are not available in the ONTAP file but
are cited by ERIC accession number at the conclusion of this discussion.
Formulation #2 in which high recall is the goal retrieves all the relevant
items In the ONTAP.file; however, nearly one out of every two citations is
139
non-relevant. Moreover, the formulation is costly as two lengthy Search Saves
must be processed, The results of the search are summarized below:
Formulation
Number of citations found = 59
Recall Number of relevant citotiNumber of citations in ins
and 28 (1007, relevant documents-et 28 retrieved)
Number of relevant (44% of retrieved doc-ions foundPrecision =.Number of r t:itluiia found 59 cuments are relevant)
The purpose of the Briefsearcli, in this case, is to perform an efficient
and rapid survey of the rile in order to determine the appropriate approach to
search strategy. It functions adequately, in this role; it could also he em-,
ployed as the first portion of a citation pearl growing approach in which a few
relevant citations are needed in order gather additional search terms. The
formulation itself, as three search statements, recalls nearly half the relevant
items' in the file. Its recall and precision ratios are given below:
Briefsearch
Number of citations found = 18
Recall
Precision
Number of relevant citations _o -d 12 (43% relevant documentsNumber of citations in answer set 28 retrieved)
Number of relevant citations found 12 (66% of retrieved doe-Number of citations found 18 uments are relevant)
Both formulations #3 and #4 require 'nimom of online connect time. They
are comprised of heavily posted search terms/descriptors. Precision varies little
between the searches but nearly 20% more relevant items are retrieved by formula--tion #3 which employs free text searching.
Formulation Free text searching)
Number of citations found = 30
Recall = Number of relevant citations found 21Number of citations in answer set 28
Precision 'umber of relevant citations found 21
Number of citations und
(75Z relevant documentsretrieved),
(70% of retrieved doc-30
uments are :elevant)
Formulation #4. (Controlled vocabulary and title searching
Number of citations found = 22
Recall Number of elevant citations found 16Number of (57% relevant documentscitations in answer set 28
retrieved)
Precision s Number of relevant citationsNumber of citations found
ound 16-(73% of retrieved doc-22-uments are relevant)
Formulation #4 achieves the highest precision; however, formulation #3
results arly the same precision with an improveMent in recall. In the
Briefsearch, two out of three retrieved items are relevant, and it is a low
cost formulation as little online connect time is necessary.
Keeping in mind the search objectives, have you developed a formulation
which satisfies these objectives? Three of the four formulations given above
may satisfy the requestor needs; which formulation including your own do you
prefer? Determining the best formulation is left to you, the reader and
searcher.
Relevant citations:
EJ 120892 EJ 119492 Ej 1194774 ,EJ 119468 EJ 119344 EJ 119342
EJ 119253 EJ'115941 EJ 112580 EJ 112578 EJ 112577 EJ 112576
EJ 111026 EJ 109399 EJ 109374 EJ 109371 EJ 109369 EJ 109368
EJ 109367 EJ 107426 EJ 107387 EJ 105970 EJ 104675 EJ 104296
EJ 103429 ' EJ 102662 EJ 102642 EJ 102322'
141
APPENDIX
COMMON SEARCH FACETS/SEARCHSAVE FORMULATIONS
When searching the ERIC fife, a number of concepts occur frequently as
facets of search topics such as
1. Grade school level
2. Population groups
3. Types of libraries
4. Geographic areas
5. Miscellaneous topics
This appendix contains search formalatl, covering three of the five
general areas listed above. The formulations have been constructed according
to the Building Block approach to search strategy; in almost every topic, a
single concept is presented. Search logic is fairly simple; the concept, its
variat forms and synonymous terms are individually SELECTed and all results
CONEINEd by the OR operator: High recall is the primary objective.
Since search terms have been limited to those having postings, it is
possible that some terms have been overlooked which may have postings in the
future; the formulations can be modified to meet specific needs or used as
given. It is suggested that searchers store any number of these searches by
using the DIALOG Search Save feature which is illustrated below in a search
for:- Junior High School (grades 7-9).
E SEARCH FACET: Junior Hi h School
1,-S GRADE 7,
S GRADE 8
3 S GRADE 9
CON JsIT AND DESCRIPTORSlldenti
GRADE 7
GRADE -8
GRADE 9
All 13 Search Saves in this section and the Search Save for WOMEN (Section II,VSearch topic #I) have been selected with permission from Selected Search
outinesines for Searchin the ERIC Data Base with the Lockheed Informationrvite, hy Charles P. Bourne, Barbara Anderson and Jo Robinson.
Berke y, CA: institute of Library Research, University of California, 1977.(ED 140870)
143
SAMPLE SEARCH FACET cont,) COMMENTS ANT DECRIPTO_ Identifiers
4 S JUNIOR(W)HIGH
5 S MIDDLE(W)SCHOOL
6 S MIDDLE(W) SCHOOLS
7 C 1-6/OR
END/SAVESERIAL#1JHU
23MAY78 8:52:56 USER 4111 COMBINing them in an OR relationship,$0.00 0.046 HRS FILE201 13 DESCRIPTORS$0.37 TYMNET$0.37 ESTIMATED TOTAL COST
JUNIOR HIGH SCHOOL ROLE, JUNIOR HIGH
.SCHOOL STUDENTS, JUNIOR HIGH SCHOOLS/3
multi-term Junior High Identifiers
MIDDLE SCHOOLS
After typing the search statements and
the searcher types END/SAVE and con-
eludes the Search Save flie rou-
tine is locked into the system when
the system assigns a number to
the strategy and prints it as part
of its response to the END/SAVE
command. In this case, serial num-
ber 1JHU is assigned. To recall and
process the strategy, type .EXECUTE,
THU
Comments concerning the search formulations are given alongside the strategy
as demonstrated above; also mentioned are descriptors (in CAPITAL letters) and
identifiers included in the formulation according fo the Thesaurus of ERIC
Descriptors, 7th edition, 1977, and the ERIC Identifier Usage Reort, 1974. The
approximate processing time and cost for-EXECUTing a Search Save is given at the
end of every formulation.
144
Portions of Search Saves can be executed using the EXECUTE STEPS command.
In this way, separate sets for every search statement are created rather than a
single set for the final statement of the Search Save. tiding the above example,
the following example demonstrates how the EXECUTE STEPS function is employed to
perform statements 1 to 3 of the Search Save.
SAMPLE SEARCH FACET (cont. COMMENTS
.EXECUTE STEPS 1JHU/3 In-this example, only statements 1 to3 are performed. The operationcannot begin after the first statement of the Search Save. If the1 91 S GRADE 1command does not contain the laststep number, e.g. .EXECUTE STEPS
2 103 S GRADE 71JHU, the entire Search Save isperformed and intermediary results
3 72 S GRADE 3 obtained.
At, the conclusion of the EXECUTE STEPS command, date and cost information are not
given. The response looks the same as if the searcher had SELECTed the
search phrases and terms manually and not employed the EXECUTE STEPS function.
Search Saves can be exchanged across DIALOG passwords. *A second user can
execute the full Search Save on Junior High Schools in the following way:
.EXECUTE 1JHU/USER 4111
If the second user wishes 'that intermediary sets be created from the exchanged
Search Save, the Save is performed in the following way!:
.EXECUTE STEPS 1JHU/USER 4111
If any readers develop such Search Saves, we at ERIC/IR would be glad to
serve as a clearinghouse for them. Just send us a copy and we will put them
from time to time in the ERIC/IR newsletter
*For a more detailed description of exchanging Search Saves, see "What's New onthe System ?" Chronolog 6, 2 (Feb. 1978): 4.
145
PRIM EDUCATION
Search Facet # 1 : Primary educa
FORMULATION
1 S EARLY(4)CHILDHOOD
2 S GRADE 1
3 S GRADE 2
4 S GRADE 3
5 S KINDERGART?
6 S PRIMARY(W)EDUCATION
7 S PRIMARY(W)GRADE
8 S
9 S
10 S
11 S
12 S
13 S
14 S
15 C
PRIMARY(W)GRADES
PRIMARY(w)PROGRAM
PRIMARY(W)PROGRAMS
PRIMARY(W)SCHOOL
PRIMARY(W)SCHOOLS
PRIMARY(W)SYSTEM
PRIMARY(W)SYSTEMS
1-14/OR
on (grades K-3)
COMMENTS AND DESCRIPTORS /Identifiers
EARLY CHILDHOOD, EARLY CHILDHOOD EDUCATION/
13 multi-term Early Childhood Identifiers
Entered as assigned index terms in order to
lessen possibility of false drops;
included descriptor, GRADE 1
GRADE 2
GRADE 3
KINDERGARTEN, KINDERGARTEN CHILDREN/
7 multi-term Kindergarten Identifiers
PRIMARY EDUCATION/Primary Education Project-
Since the truncation feature is not
currently operational in statements using
word proxiMity, PRIMARY(W)GRADE and its plural
PRIMARY(W)GRADES must be entered in two steps
PRIMARY GRADES
UNGRADED PRIMARY PROGRAMS
WONGRADED PRIMARY SYSTEM
Approximate '.075 hrs. 1 file 1.
147
ELEMENTARY EDUCATION
Search Facet Elementary Education (grades k-B)
CON} TS AND DESCRIPTORS /IdentifiersFORMULATION
1
2
S
S
EARLY(W)CHILDHOOD
ELEMENTARY(1W)EDUCATION
EARLY CHILDHOOD, EARLY CHILDHOOD EDUCATION,
14 multi-term Early Childhood identifiers
ELEMENTARY EDUCATION, ELEMENTARY SECONDARY
EDUCATION/Elementary Education Voucher.
Demonstration, 16 multi-term variations of
Elementary Secondary Education Act
3 S ELEMENTARY(W)GRADE
4 S ELEMENTARY(W)GRADES ELEMENTARY GRADES
5 S ELEMENTARY(W)PROGRAM
6 S ELEMENTARY(W)PROGRAMS UNGRADED ELEMENTARY PROGRAMS /Elementary
Programs
7 S ELEMENT_ Y (W)SCHOOLELEMENTARY SCHOOLI-COUNSELING, E.S. COUNSELORS,
E.S. CURRICULUM, E.S. GUIDANCE, E.S. LIBRARIES,
E.S.. MATHEMATICS, E.S. ROLE, E.S. SCIENCE,
E.S. STUDENTS, E.S. SUPERVISORS, E.S. TEACHERS/
Elementary School, 11 multi-term Elementary
School identifiers
S SELEMENTARY(W)SCHOOLSELEMENTARY SCHOOLS, BIRACIAL ELEMENTARY
SCHOOLS, CATHOLIC ELEMENTARY SCHOOLS
9 S GRADE 1 Word proximity, e.g. GRADE(W)1, is not used
when searching for grades"because the
possibility exists for retrieviytg a sub-
stantial number of false drops; GRADE 1
10 S GRADE 2 GRADE 2
11 S GRADE 3
12 S GRADE 4 GRADE 4
148
Search Facet (coat.)
FORMULATION
13 S GRADE 5
14 S GRADE 6
15 S GRADE 7
16 S GRADE 8
17 S INTERMEDIATE(W)GRADE
18 S INTERMEDIATE(W)GRADES
19 S INTERMEDIATE(W)SCHOOL
20 S INTERMEDIATE(W)SCHOOLS
21 S KINDERGART?
22 S PRIMARY(W)EDUCATION
23 S PRIMARY(W)GRADE
S PRIMARY(W)GRADES
25 S PRIMARY(W)PROGRAM
26 S PRIMARY(W)PROGRAMS
27 S PRIMARY(W)SCHOOL
28 S PRIMARY(W)SCHOOLS
-
29 S PRIMARY(W)SYSTEM
30 C 1-15/OR
24
31 C 16-29/OR
32 C 30-31/OR
A- roximate Process in
COMMENTS AND DESCRIPTORS
GRADE 5
GRADE 6
GRADE 7
GRADE 8
INTERMEDIATE GRADES
Intermediate School 201
Intermediate Schools
Identifiers
KINDERGARTEN, KINDERGARTEN CHILDREN
Kindergarten, Kindergartner, Kindergartners,
8 multi-term Kindergarter Identifiers
PRIMARY EDUCATION /Primary'' Education Project
PRIMARY GRADES
UNGRADED PRIMARY PROGRAMS
NONGRADED PRIMARY SYSTEM
COMBINing search statements in ranges
is limited so,the COMBINE operation is
repeated three times in this formula-
tion.
.182 hrs. ($4.55) in file 1.
149
arch Fac
FORMULATION
1 S COLLEGE(W)BOUND
SECONDARY EDUCATION
Secondary Education (grades 9-12)
COMMENTS AND DESCRIPTORS! Id iers
COLLEGE BOUND STUDENTS/College Bound Program,
College Bound Programs
COLLEGE PREPARATION
GRADE 9
GRADE 10
GRADE 11
GRADE 12
2 S COLLEGE(W)PREPARATION
3 5 GRADE 9
4 5 GRADE 10
5 5 GRADE 11
6 $ GRADE 12
7 S HICHMSCHOOL12 multi-term HIGH SCHOOL Descriptors, 2
multi-term JUNIOR HIGH SCHOOL Descriptors/
Junior ,High School, 2 multi-term Junior
High School Identifiers
8 S HIGH(W)SCHOOLSHIGH SCHOOLS, 7 multi-term HIGH SCHOOLS
9 S NONCOLLEGE(W)PREPARATORY
10 S PRECOLLEGE
-11Sp-TRE(W)COLLEGE
12 S PREP(W)SCHOOL
J3 $ PREPARATORY(W)SCHOOL
14 S PREPARATORY(W)SCHOOLS-
ra SECONDARY(w)EDUCATION
16 S SECONDARY(W)GRADE
17 S SECONDARY(W)GRADES
150,
Descriptors, JUNIOR HIGH SCHOOLS
NONCOLLEGE PREPARATORY STUDENTS
Precollege Centel
Pre College Centers
SECONDARY EDUCATION,2 multi-term SECONDARY
EDUCATION Descriptors/Secondary Education,
Secondary Education Act, Elementary Secon-
''clary Education
SECONDARY GRADES
Search Facet #3 (cont.)
FORNITIATION
18 S SECONDARY HOOL
19 S SECOND -Y(W)SCHOOLS
20 C 1-19/OR:.
CO TS AND DESCRIPTORS /Identifiers
5 multi-term SECONDARY SCHOOL Descriptors/
Secondary School, 6 multi-term Secondary
School Identifiers
SECONDARY SCHOOLS, EIRACIA1 SECONDARY SCHOOLS
Appro%inat Frocossin= Time: .170 hrs. 4.25) in file 1.
151-
152
CHILDREN
Search Facet #4: Children
FORMULATION
1 S ADOLESCEN?
2 S BOY
3 S BOYS
4 S BOY'S
5 S CHILD?
COMMENTS AND DESCRIPTORS/Identifiers
ADOLESCENCE, ADOLESCENT LITERATURE,
ADOLESCENTS/5 multi-term Adolescent
Identifiers
Boy Girl Identity Task, Boy Scouts,
3 multi-term Boy Identifiers
Entered without truncation because of
possibility of returned false drops
CHILD ABUSE, 22 other multi-term CHILD
Descriptors, CHILDHOOD ATTITUDES,-5
other multi-term CHILDHOOD Descriptors
CHILDREN. 19 other multi-term CHILDREN
Descriptor:-, CHILDRENS ART, CHILDRENS
BOOKS, CHILDRENS GAMES, CHILDRENS
LITERATURE/41 multi-term Child
Identifiers, 2 multi-term Children
Identifiers, 13 multi-term,Children's
Identifiers
6 S GIRL? GIRLS CLUBS/ 3 ulti-terM Girl Identifiers
7 S JUVENILE? JUVENILE COURTS, JUVENILE GANGS, 10 Multi-
term Juvenile Identifiers
8 S TEENAGE? TEENAGERS/3 multi-term Tennage.Identiters
9 S YOUTH? YOUTH, 16 multi -term YOUTH Descriptors/
29 multi-term Youth Identifiers
10C L-9/OR
Approximate Process Time: .030 hrt. SC I5 in file,'
)
Search Facet l 4: cont.)
TION
If the search were o cover very young children, then you might want
add the following terms.after the search save is EXECUTEd.
COMMENTS AND DESCRIPT ©RS /Identifiers
2.S BABY
3 SBABIES
INFAN?
S TODDLER?
Baby Talk
INFANCY, INFANT BEHAVIOR, INFANT MORTALITY
INFANTS, PREMATURE INFANTS/Infant
Care, 19 other multi -term Infant-
Identifiers, Infantilization,
Infantile Amnesia
Approximate Proce--in Time for Search Save Plus Additional
.074 hrs. .85 in file 1.
153
A ERI N INDIANS_
Search Facet #5: American Indians (general search)
If the objective is a high recall search, then consideration
should be given to for
specific tribal, names.
-ulating a Search Save which includes many
This routine is,comprisei of general .terms
applied -to the large population group of American Indians and does
cover individual tribal names.
FORMULATION
1 S ALASKA(W)NATIVE
2 S ALASKA(W)NATIVES
3 S ALASKAN(W)NATIVE
4 S ALASKAN(W)NATIVES
5 S AMERINDIAN?
6 S AMERICAN(W)INDIAN
7 S AFRICAN(W)INDIANS
8 S NATIVE(W )ANERICAN
S NATIVE(T)AMERICANS
10 S RESERVATIONS (INDIAN)
11 S TRIBAL
12 S TRIBE?
13 C 1-12/OR
AF.roximate P- T
COMMENTS AND DESCRIPTORS/Identifiers
Alaska Native Needs Assessment
ALASKA NATIVES
Alaskan Native Needs Assessment
Alaskan NativeS
Amerindians
not
AMERICAN INDIAN CULTURE, AMERICAN INDIAN
LANGUAGES /American Indian, 8 multirermAmerican Indian Identifiers
AMERICAN INDIANS, NONRESERVATION AFRICANINDIANS
Native American Administrator Program,
Native American Studies Programs
Native Americans
RESERVATIONS (INDIAN); The descriptor for
RESERVATIONS is used here rather than'-the free
text term since RESERVATIONS has a number
of meanings, all differing from one another.
4 multi-term Tribal Identifiers'
TRIBES'
.030 hours ($0.75) in file
AUDIOVISUAL AIDS
Search Facet #6: Audiovisual Aids
FORMULATION
1 S AUDIO( VIDEO
2 S AUDIOVIDEO?
3 S AUDIO(W)VISUAL
4 S AUDIO(W)VISUALS
5 S AUDIOVISUAL?
6 S AV
7 S CARTOON?
8 S CARTRIDG?
9 S CASSETTE?'
10 S CENTRAL(W)SOUND
11 S COLOR(W)PRESENTATION
12- 5 COLOR(W)PRESENTATIONS
13 S DIAL(W)ACCESS
14 S DISPLAY(W)PANEL
15 S DISPLAY(W)PANELS
16 S DOCUMENTAR?
CO TS AND DESCRIPTORS/Ide_
AUDIO VIDEO LABORATORIES/3 ulti-term
Audio Video Identifiers
9 multi-term Audio Visual Identifiers
7 multi-term AUDIOVISUAL Descriptors/10multi-term Audiovisual Identifiers
The truncation capability is not used in
this search statement because terms such
as avenue, aviation, average, or avocation,
would be retrieved; included identifier;
AV Communication Review
CARTOONS
3 multi. -term CASSETTE or. CASSETTES
Descriptors/Cassettes, 5 multi7term
Cassette Identifiers
CENTRAL SOUND SYSTEMS
COLOR PRESENTATION
DIAL ACCESS INFORMATION SYSTEMS /Dial
Access Information Retrieval System,
Dial Access Laboratories
DISPLAY PANELS
DOCUMENTARIES/Documentary Drama, Documentary
History, Documentaries'
Search Facet #6 (cont.)
. FORMULATION
17 S EXHIBIT?
18 S EXPOSITION?
19 S FELT(W)BOARD
20 S FELT(W)BOARDS
21 S FILM?
22 S GRAPHIC(W)AID
23 S GRAPHIC(W)AIDS
24 S GRAPHIC(W)ART
25 S GRAPHIC(W)ARTS
26 S GRAPHICS
27 S INSTRUCTIONAL(W)AID
28 S INSTRUCTIONAL(W)AIDS
29 S INSTRUCTIONALMMEDIA.
30 S LANGUAGE(W)AID
31 S LANGUAGE(W)AIDS
32 S LANGUAGE(W)LAB
33 S LANGUAGE(W)LABS
34 S LANGUAGE(W)LABORATORY
COMMENTS AND DESCRIPTORS/Iden
EXHIBITS
EXPOSITIONS
FILMS, 7 multi-term FILM Descriptors,
4 multi-term FILMS Descriptors, FILM-7
STRIP- PROJECTORS, FILMSTRIPS/31 multi-
term Film Identifiers,Pilmic Stylesi
Filmlists, Filmographies, Filmography,
Filmography of Films about Movies,
Films Incorporated, Filmstrip Viewers
GRAPHIC ARTS /Grapl c Arts Industries
COMPUTER GRAPHICS, ENGINEERING GRAPHICS/
Graphics Expression Reading improvement
System, Graphics Expression System
INSTRUCTIONAL AIDS/Instructional Aids
INSTRUCTIONAL MEDIA/Instructional Media Centers
LANGUAGE AIDS
_GUAGE LABORATORY EQUIPMENT, LANGUAGE,
LABORATORY'USE/Language Laboratory,
Language Laboratory Directors, Language
Laboratory EquipMent
-arch Facet if 6 cont.)
FORMULATION
_ LANGUAGE( LABORATORIES
36 S LEARNING(W)LABORATORY
37 S LEARNING(W)LABORATORIES
38 S MICROPHONE?.
39 S MULTI(W)MEDIA
40 S MULTIMEDIA
41 S OVERHEAD(W)PROJECTOR
42 S OVERHEAD(W)PROJECTORS
43 S PHONODISC?
44 S PHONOGRAPH?
45-S PHONOTAPE
46 S PHOTOGRAPH?
47 S POSTER?
48 S PROJECTION E UIPMENT
49 S RADIO
50 S RADIOPHONIC
51 S RADIOS
52 S REALIA?
53 S SCREENS (DISPLAYS)
54 S SLIDE(W)PROJECTOR
55 S SLIDE(W)PROJECTORS
COMMENTS_AND DESCRIPTORS/Identifiers-
LANGUAGE LABORATORIES
Learning Laboratory
LEARNING.LABORATORIES
MICROPHONES
Multi Media Techniques, Multi. -Media Treatise...
MULTIMEDIA INSTRUCTION/5 multi-term
Multimedia Identifiers
Overhead Projector Animation System
OVERHEAD PROJECTORS_
PHONOGRAPH RECORDS, LANGUAGE RECORDS
(PHONOGRAPH)/Photographic Representation,
Phonographs
PHONOTAPE'RECORDINGS/Phonotape Records
PHOTOGRAPHIC EQUIPMENT, PHOTOGRAPHS,
PHOTOGRAPHY /-2 multi-term Photograph
Identifiers, 3 multi-term Photographic
Identifiers
PROJECTION:EQUIPMENT
RADIO, EDUCATIONAL RADIO,_RADIO TECHNOLOGY/
17 multi-term Radio Identifiers,
Radiophonic Teaching
REALIA
SCREENS- ( DISPLAYS)
Seareh Facet cont.)
FORMULATION
56 $ SLIDE(W)TAPE
57 S SLIDE(W)TAPES
58 S SLIDES
59 S TAPE(W)RECORDER
60 S TAPE(W)RECORDERS
61 S TAPE(W)RECORDING
62 S TAPE(W)RECORDINGS
COMMENTS AItiiD DESCRIPTORS/Identifiers
Slide Tape Sequences
DES
TAPE RECO
TAPE RECORDINGS, VIDEO TAPE RECORDINGS
63 S TELEVISfON?TELEVISION, 21 multi-term TELEVISION
Descriptors/16 multi-term Television
Identifiers
64 S TRANSPARENC? TRANSPARENCIES
65 S TVIdentifiers: TV,,,TV College,
TV College (Chicago)
66 S TV'S
67 S VIDEO?
68 S VIEWE
69 S VIEWING(W)CENTER
70 S VIEWING(W)CENTERS
71 S VISUAL(W)AID
72 S VISUAL(W)AIDS
4 multi-term VIDEO Descriptors/11 multi-
term Video Identifiers, Videographics
SyStems, Videopublishing, Videorecorda;
Videosonic Opticon Application, Videosonic
Teaching Machine, Videotspejarticipation
-System, Videotape Program Servic
VISUAL AIDS/Visual Aids
Search Facet #6 cont.)
'FORMULATION
73 S VISUALS
74 C 1-25/OR
75 C 26-50/OR
76 C 5143/OR
77 C 74-76/OR
CO NTS AND DESCRIPTORS
cimateinmeArol: .261 hours. ($6.5
Ideitif or
n file 1.
EVALUATION (HIGH RECALL)
Search Facet 117: EVALUATION
This is a high recall search for the concept of Evacuation,
Full use of free text searching ancrtruncation is made; over
150,000 postings are retrieved Whert-p essing this search in
DIALOG's full ERIC. Search Save 118, on the other hand, is a
high precision formulation for Evaluation.
FORMULATION
1 S ANALY?
2 5 A2PRAIS?-
3 S ASSESS?
4 S COMPAR?
5 S COST
6 S COSTS
COMMENTS AND DESCRIPTORS /Identifiers
30 Multi -term ANALYSIS Descriptors, ANALYSTS
ANALYTIC GEOMETRY, ANALYTICAL CRITICISM!'
5 multi -term Analysis Identifiers
8 multi-term Analytical Identifiers,
Analyticity, Analytical Methods,
Analytic Philosophy, Analytic Rotational
Methods
PROPERTY APPRAISAL /Appraisal of Personal
Growth
ASSESSED VALUATION, EDUCATIONAL ASSESSMENT-,
INFORMAL ASSESSMENT; NEEDS ASSESSMENT,
PERSONALITY ASSESSMENT/3 multiterm
Assessing Identifiers, Assessment,
7 multi-term Assessment Identifiers
COMPARATIVE ANALYSIS, COMPARATIVE EDUCATION
COMPARATIVE STATISTICS, COMPARATIVE TESTING,
AUDIO ACTIVE COMPARE LABORATORIES,
EDUCATIONAL STATUS COMPARISON/Comparability,
Comparable Cost Information, 23 multi-term
Comparative Identifiers, Comparing
Reading Approaches
COST EFFECTIVENESS, COST INDEXES/Cost
Accounting, 9 multi -term Cost Identifiers
COSTS, 9 multiterm COSTS Descriptors/
Costs of Schools Training and Education
Search Facet TT cont.)
FORMULATION
7 S CRITER?
8 S EFFECT?
9 S EVALUAT?
10 S FEASIB?
11 S IMPACT?.
12 S MEASUR?
13 S PERFORMANCE?
CO_ NTS AND DESCRIPTORS /Identifiers
CRITERIA, 6 multi-term CRITERIA Descriptors,
CRITERION REFERENCED TESTS/Criteria of
Success In English Test, 5 multi-term
Criterion Identifiers
.PRIMACY EFFECT, EFFECTIVE TEACHING', 4 multi-
term EFFECTIVENESS Descriptors, 4 multi-
term EFFECTS Descriptors/Effectance
Motivation, Effectively Influencing
Political Decisions, 'Effectiveness,
5 multi-term Effective Identifiers,
Effects of Daylight Savings Time.
EVALUATION, ADMINISTRATOR EVALUATION, 26
other multi -term EVALUATION Descriptors,
EVALUATIVE THINKING/Evaluation,
10 multi-term Evaluation Identifiers,
3 multi-term Evaluative Identifiers
Evaluators
FEASIBILITY STUDIES
IMPACT,-7 multi-term Impact Identifiers
MEASUREMENT, 6 multi -term MEASUREMENT
Descriptors, VISUAL MEASURES/6 multi-
'term Measurement Identifiers, Measure
of Sampling Adequacy, 4 multi-.
Measures Identifiers,,Measuriag
Instruments
PERFORMANCE, 11 multi-term PERFORMANCE
Descriptors/19 multi-term Performance
Identifiers
Search Facet 117 cont.)
FORMULATION
14 S RELIAB?
S SUCCESS
16 S SUCCESSES
17 S SUCCESSFUL
18 S VALID?
19 C 1-8/OR
COMMENTS AND DESCRIPTORS /Identifiers-
RELIABILITY,' TEST RELIABILITY /Reliability
Analysis Center
SUCCESS FACTORS
VALIDITY, PREDICTIVE VALIDITY, TEST
VALIDITY/Valid Confidence Validity
Research, Validity Study Service
Auroximate Processing Time: .068 hrs. ($1.70) n file
EVALUATIO
Search Face
(HIGH PRECISION)
EVALUATION
This is a high precision search formulation for the concept of
Evaluation. It is restricted to searching only the identifier 'title,
and descriptor fields of the ERIC r
all in this formulation in order
retrieving false drops.
ord. Truncation is not used a
to eliminate the possibility of
Because searching is restricted tl three of
the 7 possible subject-conveying fields in the ERIC record, the
of Search Save #8 is only half the siz
FORMULATION
1 S ANALYSIS DE,ID,TI
2 S ANALYSES/DE,ID,TI
3 S ANALYTIC/DE,ID,TI
4 S APPF,AISAL/DEJJ,TI
5 S APPRAISING/DE ID,TI
6 S ASSESSING/DE,ID,TI
7 S ASSESSMENT/DE,ID,TI
8 S COMPARATIVE/DE,ID,TI
9 S COMPARING/DE,ID,TI
10 S COMPARISON/DE,ID,TT
11 S COSTS/DE,ID,TI
of the final set. In #7.
result
COMMENTS AND DESCRIPTORS/Identifiers
30 multi-term ANALYSIS Descriptors/5 multi-,
term Analysis Identifiers
ANALYTIC GEOMETRY
PROPERTY APPRAISAL/Appraisal of Personal
Growth'in Teamwork
3 multi -term Assessirig Identi er
EDUCATIONAL ASSESSMENT, INFORMAL ASSESSMENT,
NEEDS ASSESSMENT, PERSONALITY ASSESSMENT/
Assessment, 7 multi-term Assessment
Identifiers
4 multi-term COMPARATIVE Descriptors/23
multi-term Comparative Identifiers
2 multi-term Comparing Identifiers
COSTS, 9 multi-term COSTS Descriptors/
Costs of Schools Training and Educatio
7
Search Facet
FORMULATION
12 S CRITERIA DEID,TI
13 S CRITERION/DE,ID,TI
14 S EFFECTIVENESS/DE,ID,TI
15 S EVALUATION / DE,ID,TI
16 S EVALUATIVE/DE,ID,TI
17 S FEASIBILITY/DE,ID TI
18 S IMPACT/DE,ID,TI
19 S MEASUREMENT/DE,ID,TI
20 S PERFORMANCE/DE,ID,TI
21 S RELIABILITY/DE,ID,TI
22 S SUCCESSFUL / DE,ID,TI
2355 VALID/DE,ID,TI
24 S VALIDITY/DE,ID4TI
25 C 1-24/OR
COMMENTS AND DESCRIPTORS/Identifiers
CRITERIA,,6 multi-term CRITERIA Descriptors /_'
Criteria of Success in English Test
CRITERION REFERENCED TESTS/5 multi-term
Criterion identifiers,
4 ultiterm EFFECTIVENESS Descriptors( )
EVALUATION, 26 multi-term EVALUATION
Descriptors/Evaluation,'10 multi-term
Evaluation Identifiers
EVALUATIVE THINKING/8 multi-term Evaluative
Identifiers
FEASIBILITY STUDIES
MEASUREMENT, 6multi-term MEASUREMENT
Descriptors/6 multi-term Identifiers
PERFORMANCE, 11 multi-term PERFORMANCE
Descriptors/19 multi-term Performance
identifiers
RELIABILITY, TEST RELIABILITY /Reliability
Analysis Center
Valid Confidence
VALIDITY, PREDICTIVE VALIDITY, TEST
VALIDITY/Validity Research, Validl
Study Service
roximate Processing Time: .031 hrs. ($.78) in f
164
JUNI COLL EGE
Search Facet #9: Junior College
The to NIOR COLLEGE and COMMUNITY COLLEGE are searchable in
subject- conveying fields except corporate source, sponsoring agency and 'de-
scriptive note in order to exclude its sponsored and /or authored by a
specific educational institution.
FORMULATION COI NTS AND DESCRIPTORS Identifiers
1 S ASSOCIATE(W)DEGREE
2 S ASSOCIATE(W)DEGREES
3 S COMMUNITY(W)COLLEGE/DE,
ID,TI,AB.
4 S COMMUNITY(W)COLLEGES
'5 S GRADE 13
6 S GRADE 14"
7 S JUNTOR(W)COLLEGE/DE,IT;
TI,AB
8 S JUNIOROOCOLLEGES
9 S POSTMSECONDARY
10 S POSTSECONDARY
11 C 1-10/OR
ASSOCIATE DEGREES
8 multi-term Community College Ident
COKMUNI _ COLLEGES
GRADE 13
GRATE 14
JUNIOR COLLEGE LIBRARIES, JUNIOR COLLEGE
STUDENTS/6 multi-term Junior College
Identifiers
JUNIOR COLLEGES
POST SECONDARY EDUCATION/Post Secondary
Schools
°oximate Processin Time: .049 hrs. ($1.23) in file 1.
Search Facel 0
HIGHER EDUCATION
Higher "Education
The terms COLLEGE COLLEGES, HIGHER EDUCATION, GRADUATE SCHOOL, UNIVERSITY,'
and UNIVERSITIES are searchable ubject-conveyintfields except corporate-,
source, sponsoring agency and descriptive note fields:' This-reduces the number
of false drops which may occur when a particular educational institution is
responsible for the document
FORMULATION
as a corporate author or. sponsor.
-1 S COLLEGE/DE,ID,TI,_
2 S COLLEGES/DE,ID,TI,AB
3.S cot EGIA?
4 S DOCTORAL
5 S GRADUATE(W)SCHOOLIDE,ID,TI,AS
6 S GRADUATE(W)SCHOOLS
7 S GRADUATE(W)STUDENT
8 S GRADUATE(W)STUDENTS
9 S GRADUATEMSTUDY
10 S GRADUATE(V)STUDIES
11 S HIGHERMEDUCATION DE,IDT
S POST(W)SECONDARY:
13 S FOSTSECOND_
14 :z UNDERGRADUATE?
COMMENTS AND DESCRIPTORS /Identifiers
COLLEGE ADMINISTRATION,- 40_other multi term
COLLEGE Descriptors/College, over 60 ,t .
multi - term. College Identifiers
-COLLEGES,
tors
-Collegial System, Collegial Teams, four-
Aulti-term Collegiate Identifiers.
DOCTORAL DEGREES, POST 0P0TORWEDUrATION;-
DOCTORAL PROGRAMS, DOCTORALTHESES/-
Doctoral Thesa,cpectoralThesid-,
Graduate School Foreign fanguage Test,
multi-term_COLLEGE S,_D scrip-,
Gradudte Schoolof Education
Graduate Schools, 2 multi-term. G
schools identifiers
UATE STUDENTS
GRADUATE STUDY
I HER EDUCATION/32 mull
Education Identifiers
POST SECONDARY EDUCATION/Po
. Schools
rm Higber
Secondary
UNDERGRADUATE STUDENTS',,,UNDERGRADITATE STUDY /- _
5= multi7term Undergraduate Identifiers:;a
Search Facet 10 c-nt.)
FORMULATION
15 S UNIVERS /DE,ID,TI,AB
16 S UNIVERSITIES/DE,ID,TI,AB
17 C 1-16/OR
CO. NTS Ah D DESCRIFTO dentif
Three multi -term UNIVERSITY Descriptors/
University, over 100 multi-term
Identifiers
UNIVERSITIES, multi -term UNIVERSITIES
Descriptors /3 multi-term Universities
Identifiers
Approximate Processiss_Tilt: .076 hrs. ($1.90} file
167
HANDICAPPED
Search Facet 1 /11: Mentally, neurologically and emotionally handicapped
FORMULATION
1 S APRASI?
2 S AUTISM
3 S BRAIN(W)DAMAGE
4 S BRAIN(J)DAMAGED
5 S BRAIN(W)DYSFUNCTION
6 S BRAINMINJURED
7 S BRAIN(W)INJURY
B S DEVIANT(W)BEHAVIOR
9 S DYSLEX?
10 S HANDICAP?
11 S HYPEFZACTIV?
12 S MALADJUSTED(W)CHILD
13 S MALADJUSTED(W)CHILDREN
14 S MALADJUSTMENT
15 5 MONGOLISM
16 S MONGOLOID?
17 S NEUROS?
18 S NEUROTIC?
19 S PROBLEM(W)CHILD
20 S PROBLEM(W)CHILDREN
21 S PSYCHOS?
22 S PSYCHOTIC?
23 S RETARDED
24 S SCHIZOPHRENI?
25 C 1-24/OR
26 S BEHAVIOR?
27 S EMOTIONAL?
28 S LEARNING
168
COMMENTS AND DESCRIPTORS /Identifiers
APHASIA/Asphasic
AUTISM
Brain Damage
Brain Damaged
MINIMALLY BRAIN INJURED
SOCIALLY DEVIANT BEHAVIOR
DYSLEXIA
HANDICAP DETECTION, HANDICAPPED, 21 multi -term
HANDICAPPED DESCRIPTORS, 3 multi-term
HANDICAPS DESCRIPTORS
HYPERACTIVITY
MALADJUSTMTNT, EMOTIONAL MALADJUSTMENT
MONGOLISM
NEUROSIS
NEUROTIC CHILDREN / Neuroticism, Neuroticism
Scale Questionnaire
PROBLEM CHILDREN
PSYCHOSIS
PSYCHOTIC CHILDREN/Psychotic Impatient Profile
3 multi-term RETARDED Descriptors/ 4 multi-
term Retarded Identifiers
SCHIZOPHRENIA/Schizophrenia
BEHAVIOR PROBLEMS/Behavior Problem
Checklist, Behavioral Problems
EMOTIONALLY DISTURBED, EMOTIONALLY DISTURBED
CHILDREN/Emotionally Disturbed Children
Learning Disability Teaching Consultant,
LEARNING DISABILITIES/Learning Disabilities
Act of 1969
Search Face 1 (cant.)
FORMULATION
29 S MENTAL?
30 S NEUROLOGICAL?
31 S PERCEPTUAL?
32 S PERSONALITY
33 C 26-32/OR
34 S DEFECT?
35 S DEFICIEN?
36 S DISABILIT?
37 S DISORDER?
38 S DISTURB?
39 S ILL
40 S ILLNESS?
41 S PROBLEM?
42 C 34-41/OR
43 C 33 AND 42
44 C 25 OR 43
COMMENTS ANDDESCRIPTORS/Identifiers
MENTALLY HANDICAPPED, CUSTODIAL MENTALLY
HANDICAPPED, EDUCABLE MENTALLY HANDICAPPED,
TRAINABLE MENTALLY HANDICAPPEDNEUROLOGICAL
DEFECTS/NEUROLOGICALLY
RANDICAPPED/Neurologically Handicapped
PERCEPTUALLY HANDICAPPED
PERSONALITY PROBLEMS
Approximate.139 hrs. 48) in file 1.
169
PRESCHOOL
Search Facet #12: Preschool
COMMENTS AND DESCRIPTORS/IdentifiersFORMULATION
1 S EARLY(W)CHILDHOOD EARLY CHILDHOOD, EARLY CHILDHOOD EDUCATION/
13 multi-term Early Childhood Identifiers2 S KINDERGART?
KINDERGARTEN, KINDERGARTEN CHILDREN/ 7 multi-
term Kindergarten Identifiers3 S NURSERY(W)SCHOOL
4 S NURSERY(W)SCHOOLS Nursery Schools
5 S PREKINDERGART?
6 S PRE(W)KINDERGARTEN
7 S PRE(W)PRIMARY
8 S PREPRIMARY
9 S PRE(W)SCHOOL
10 S PRESCHOOL?PRESCHOOL CHILDREN, 9 multi-term PRESCHOOL
DESCRIPTORS/20 multi-term Preschool
Identifiers
roximate Processin Time: .03 ($0.75) In file 1.
The above formulation is rather specific to the preschool age group. If
a high recall search is desired, consideration might be given to adding the
following two terms which include the prechool age group, but also include
other groups.
11 S CHILD
12 S DAY(W)CARE
170
CHILD CARE, CHILD CARE CENTERS, MIGRANT CHILD
CARE CENTERS, CHILD CARE OCCUPATIONS,
CHILD CARE WORKERS/Child Care Center
FAMILY DAY CARE, DAY CARE PROGRAMS, DAY CARE
SERVICES/Day Care Centers, 5 multi-term
Day Care'Identifiers
APPENDIX
GUIDE TO DIALOG'SERIC DATA BASE
The database description for ERIC prepared by DIALOG for their volume
entitled, Guide o DIALOG -- Databases (Palo Alto, CA: Lockheed Information
Systems, August 1977) is reproduced here almost in its entirety (only a few
pages of Search Examples are omitted).
This description should be an aid to the reader of this manual as it
provides the most definitive exposition of the ERIC files (RIE and CIJE) as
they appear and are accessed on the DIALOG Information Retrieval Service.
171
FILE 1
DIALOG* INFORMATION RETRIEVAL SERVICEERIC
FILE DESCRIPTION
ERIC is the complete database on educational materials from the Educational Resources Information Center.It consists of two main files: Resources in Education (RIE), which is concerned with identifying the mostsignificant and timely education research reports; and Current Index to Journals in Education (CUE), an indexOf more than 700 publications of interest to every segment of the educational profession. (11 1.1)
SUBJECT COVERAGE
The ERIC database includes an enormous variety of educational information in the following general areas:(Ii 1.2)
Career EducationCounseling and Personnel ServicesEarly Childhood EducationEducational ManagementHandicapped and Gifted ChildrenHigher EducationInformation ResourcesJunior Colleges
SOURCES
Languages and LinguisticsReading and Communication SkillsRural Education) and Small SchoolsScience, Mathematics, and Environmental EducationSocial Studies/Social Science EducationTeacher EducationTests, Measurement, and EvaluationUrban Education
ERIC indexes a wealth of document types: research reports, evaluation studies, curriculum guides, lessonplans, bibliographies, course descriptions, theses, journal articles, pamphlets and other "fugitive" materials.Most of the item!, aside from journal articles, can be purchased from the ERIC Document ReproductionService (ERRS) in paper copy or microfiche. There are approximately 650 locations throughout the countryhaving collections of the ERIC microfiche, and most are open to the general public. (t 1.3)"
DIALOG FILE DATA (11 1.4)
Inclusive Dates:Update Frequency:File size:
ORIGIN
ERIC is produced by:
1966 to the presentMonthly (approximately 3,000 a month)280,000 citations, as of June 1977
National Institute of EducationEducational Resources Information CenterWashington, DC 20208
Questions concerning this file should be directed to:
User Services CoordinatorERIC Processing and Reference Facility4833 Rugby Avenue. Suite 303Bethesda, MD 20014
.Ttedernark Reg. U.S. Pat. & Trademark Office
Telephone: 301/656-9723
1.6)
FILE 1
ERICDIALOG FILE I
SAMPLE RECORDDIALOG Arr.IlioAc:11.44L
7,777!i 7.!: .. JJ4= Co; J=O==.=/71A11=..=1500... Ed.ord. A=d 0,==Jt
j,, E,t..osi ,O El Jol 1=t A1l=,=9,O= P C /CSNg, o-o J=i7 O, E0.49..0. 11:O=LA, o. -n .==-/59D C
Nl Si 2.0en2Ne OIC S 8 07086' OCC!
Pi,01 O.,. 71 Not* f= 9=== 5)8J=j0,±J I.6 St 9=CJ va
f DE 5 50 T. Off...DotJ.lj=j5 .C===,JtjJ.
C., .,=."99
b,.f.., f g east:_.itg,. ' 5=1.01
_ z, co...o= wi.. 1, 57
/NTFor
f.. Tft,,,... CC !?* P53 'a-
!DE
ID
liogot Old fo,J=C.I. ed,ca,Joft01 . to59Z''9=9. tPu.' -rs. ,'' '.@ '.....'., oh. =.0 81 Of
Cl+5,5"7 .C1.0,.. 1,0116501 2/d 90'*,=1 0. CU I ft t Q'
0,= 1510,,Jdod '0 J1J=1,,o1= t=. majO# so- roe ,, u, t==Jcollo=OrAo,.0= jr=1J-11
RETRIEVAL METHODS=
SUWECT OR TEXT SEARCHING
1 7)
_i SUFFIX FIELD NAME EXAMPLES ---1
1. N000 -Jame Imkx(171=19thl all *p.m field. 1 E CURRICULUM S CHEMISTRY
1,3 /AB lAtMreetSCIENCEMEOLICATION
S NrogaqmqcoLLEGES/A0I hotclf 000./
1-4I
/L,S I CflopmCg ?'00 AA S SERVICESIIWIEDUCATION(FIwASHINGTON/CS1-4 /DE 1 EIvicnplof IS PHYSICAL SCIENCES
1401$
/NT ' DAtonwor Amc/10 1 Identawr IS THIRTEEN0000OLLEGETV@WURRICULUMADIS
S APPENDIX/1NT1.10 /SA i SmInnoring Asemel, I S NATIONALMINST(IWIEDUCATION/SA
. I U rrl Intl= iScuromuLum(wlrgooRAmiT1
'AlsorDE.,MDF./Dr../1D'=11EAF'
PACeF I PREFIX
I II AC.1121-131-1,.1.15
146I-17
1-111
1 -18
114Lid
AU=CHChl -DT=GCIS =-
10 -RN=RNSC.
LID=
CODE SEARCHING
FIELD NAME
Leitilative Authority CodeAuthorChmemshome CodeCoutmetiOrmt NumberDocument TypeGroup Codelujourn6I NomeBuresu/Trojeci NumberRepoti NumberCO/Eporafg Source Code orSponionng Aieney CodeUpdateYmr
EXAMPLES
E ACA 5.6E ROOMER, EE CH= VTE CN =OECE DT =t1E (3C= 1111 (CURRICULUM)E IS .CIIE1977E JO. AAUPE PN. BR
RN.BE SC= FORE SC= BBBE LID= 7812E YR =74
S AC=95S BOOKER. EDWARD5 CtlS CN .0EC13.11.07567,X431S 01-c-tExPAND 0.4)S 15 =RIEAPR76S ID- INTERCOMS PN .BR-7-0167S RN.114197S 5C= FOK37812S SC= EIBB068215 = 7e045 YR =76
_ MNGPAGE SUFFIX FIELD NAME EXAMPLES
7.10 /ED Memnon Numbem foulior ED Subfile LI 7 2 D LL4IEDfF_I Amemino Number' And/or FJ &Mille L5 I 2 LETO
1-20 /AVAIL Document Armlable from MR LIMAYFUNAVAIL Decummt Non Amiable from EDRS LI4fUNAVAIL
1.20 /MAT I.4A0oe Drecriptor or Identifier L7/MAJ/MIN Minor Dmeriptor or Identifier Ln/MIN
DIRECT ACCESS
EXAMPLESL__EA(iF PREFIX FIELD NAME
1.24 1Norle (DIALOG Am.-Arson Number i T [011)507/3 D ED115307
FORMATS AVAILABLE1 PAGE. NUMBER 1 RECORD CONTENT
1.22 Formal I rDIALOO Accettion NumberFt...TIM 2 COFT.Pief. Record Except AbmiemFormat 3 - Ribltotrephic C116uOrs
[- -Formal 4 AbetruciFellla=r 3 Full Record
_t__) Pormat 6 Title_ __._
FILE 1
SEARCH OPTIONS
BASIC INDEX, If a prefix or suffix is not specified in an EXPAND or SELECT command, the commandis executed in the basic index. The basic index contains all assigned descriptors (Thesaurus terms) and identifiers,plus all meaningful individual words exoacted from the Title, Abstract, Descriptor, Identifier, CorporateSource, Descriptive Note, and Sponsoring Agency fields. Thus, all words of subject importance can be searchedwithout regard to their location within a given record_ However, the full-text capability can be used to definea search statement further by specifying the desired proximity and order of words, their occurrence in a givenfield, or by any combina.ion of these as shown in the following sections.
(11 1-9)
ABSTRACT (/AB). Abstracts for the report literature (RIE items) are generally short (less than 200-word)summaries of the document contents. Abstracts for the journal articles (CUE items) are generally brief (lessthan 50.word) annotations. Annotating is done only to enrich those articles with uninformative, non-expositorytitles.
(11 1.1©)
The use of free language terms that might occur in abstracts, titles, or identifiers is particularly useful forsearch topics in new research areas for which the Thesaurus vocabulary may not be adequate.
(11 1.11)
Acronyms may be used in abstracts although individual words generally are not abbreviated. The searchershould include the full phrase and its acronym in the strategy.
(11 1.12)
? SELECT MMPI /A8
1 130 MM AB? TYPE 1/4
EJ149928This study investigated objections to Minnesota MultiphastcPersonality Inventor's, (MMPI) items by black and white males and
females classified as high or low trust by Rotter's InterpersonalTrust Scale: Low-trust subjects objected to more items than high -trustsubjects. Race and sex did not influence number of objected to.(Author)
? SELECT MINNESOTA(W)MULTIPHASIC(WIPERSONALITY/AB2 19 MINNESOTAIMMULTIPHASIC(W)PERS
? COMBINE 1 OR 23 163 1 OR 2
Names of individuals in the text of the abstract may or may not include initials or given name with the surnameand may occur in the possessive form. Truncation can be used with the surname to encompass both forms.
(1l 1.13)
SELECT PIAGET?4 1428 PIAGET?
? TYPE 4'4
E.1152391
Describes a study to extend research of Riegel and Jahoda intochildren's spatial Conceptions, Individual testing and interviewing
FILE I
are d,scussed. Results indicate differences in rate of spatial stagesdevelopment of Swiss. Scottish and American children, althoughPia ets' original theorized stages remain invariate. (Author/AV)
CORPORATE SOURCE ( /CS). The Corporate Source field, found only in RIE items, is used to list theorganization responsible for the issuance of the document. (It is not the address of the first personal author.)Each corporate source name has been entered in a standardized form and corresponds to a specific eight-characteralphanumeric institution code searchable in the SC= field_
(li 1.14)
Searching in the /CS field will generally yield broader retrieval than the SC= field because unique codes areassigned to each division of each institution.
(IT 1.15)
Full-text methods should be used to search corporate source names. It is generally best to SELECT only themost specific parts of the name and igriure the common words with no subject content. Abbreviations maybe used. For example, to find documents issued by the various divisions of the Minnesota State Dept. ofEducation,
(Ii 1.16)
? SELECT MINNESOTAMSTATE(2W)EDUCATION/C55 154 MINNESOTA(WISTATE(201)EDUCATION
? TYPE 5
EDI34491 SOD09706Political Involvement. Commieity 'Involvement/Career Education: An
Experience-based Social Studies -Program, Volume III.Kyle, James; And OthersMinnesota State Dept. of Education, St. Paul. Div. of Vocational' and
Technical Education.; Robbinsdale Independent School District 281.Minn,
Publ. Dote: Jan 75 Not705; Many pages of the originot ovoilable. They are notEDRS Price MF-$0.83 HC ;8:Descriptor *Career E
Involvement/ Democratic Via
Instruction/ InstructionalPower/ 'Politics/ RelevanceStudies/ Social StudiesVolunteer Study Guides
e: 162p.; For related document, see SO 009nal document are copyrighted and thereforeincluded in the pagination69 Plus Postage.ducation/ Career Planning/ Communitylues/ Educational Objectives/ IndividualMaterials/ Learning ActivOties/ Political(Education)/ Secondaly Education/ SocialUnits/ Student Partkioation/ Student
DESCRIPTOR (/DE, /DE*, /DF, /DF*).. Descriptors are single-word or multiword subject terms assignedfrom the Thesaurus of ERIC Descriptors to characterize the substantive content of the original document_ Onthe average, ten descriptors are assigned to each indexed document in the RIE portion of the database, andfewer descriptors are assigned to the articles in the CUE portion of the database. The assignment of descriptorsfrom the Thesaurus is discussed in detail in the ERIC Processing Manual. The Thesaurus is reissued annuallyto incorporate vocabulary changes.
(11, 1.17)
175
FILE 1
Thesaurus terms and the relationships among them have been incorporated into the online index as an aid toDIA LOG system searchers. It is possible to view the relationships by further EXPANDing any term in thebasic index that shows postings under the "RT- heading, as illustrated below.(11 1.18)
? EXPAND REMEDIAL READINGRef
El
Index-term TydeREMEDIAL EDUCATION
Items RT
PROGRAMS12 REMEDIAL HANDWRITING--
E3 REMEDIAL INSTRUCTION-- 712 16E4 REMEDIAL MATHEMATICS-- 164 6ES REMEDIAL PROGRAMS------- 586 11E6 REMEDIAL READING- - - - - -- 889 8E7 REMEDIAL HEADING CLINICS 38 3E8 REMEDIAL READING PROGRAME9 REMEDIAL READING ,
PROGRAMS- - -- 369 4EIO REMEDIAL STRATEGIESEll REMEDIAL TEACHERS 41 2E12 REMEDIAL TEACHINGE1-3 REMEDIAL WRITINGE14 REMEDIALLY 2E15 REMEDIALS__ 3El6 REMEDIALSUMMER-[17 REMEDIATE = 62E18 REMEDIATED- 8
? EXPAND E6Ref Index-term Type Items RT .
RI -REMEDIAL READING 889 8R2 CORRECTIVE READING - N 55 1
R3 READING----- - -- - 8 22699 71R4 REMEDIAL COURSES B 84 6RS READING CENTERS---- - - - - -R 176 4
R6 REMEDIAL INSTRUCTION----R 712 16R7 REMEDIAL READING CLINICSR 38 3R8 REMEDIAL READING
PROGRAMS--------------R 369 4R9 RETARDED READERS--------R 227 5
The following term relationships are coded in the "Type" column:
1.19)
B - Broader term
N - Narrower term
R Related term
U - Use or Used for
FILE 1
Ter Ms such as El above (without postings in the basic index but with a related-term posting) are included toprovide cross references to proper Thesaurus terms. Thus, the searcher should EXPAND the E referencenumber to locate the preferred term, e.g..
T. 1.20)
? EXPAND ElRef Index -term Type Items RTRI -REMEDIAL EDUCATION
PROGRAMS------- - - - - --
R2, REMEDIAL PROGRAMS- - -U 586 11
Here the "U" indicates that he searcher should use "remedial programs" rather than "remedial educationprograms."
(II 1.21)
If the searcher is familiar with the Thesaurus, the related terms can be EXPANDed directly by enclosing theterm in parentheses, e.g.,
(II L22)
? EXPAND (LITERATURE APPRECIATION)Ref Index-term Type Items RTRI -LITERATURE APPRECIATION- 1093 4
R2 READING ENJOYMENTR3 CHORAL SPEAKING--:-----R 44 7
R4 LITERATURE -R 16515118R5 LITERATURE GUIDES-------k 158 4
Here the "U" indicates that "Iiierature appreciation" is used for or preferred over "reading enjoyment."(II .23)
Some 7'he5-4urus terms may include a scope note for further definition. Scope notes are riLLt entered when theterm is SELE.CTed, but appear in the syste'm response, e.g.,
Long
? SELECT READINESS6 2345 READINESS PREPAREDNESS TO RE
or terms with long scope no s are runcated at 42 characters as is the entry above.
(I 1.24)
(11 1.25)
Also, some Thesaurus terms may include a parenthetical word or phrase as part of the term. In such a casethe parenthetical etpression must be included in the entry to restrict retrieval to that specific term. e.g,.
1.26)
? SEIECI READINESS (MENTAL)7 122 READINES MENTAL)
(However, the- citations in set 7 would also be included in the single-word index entry in set 6.)T, 1.27)
1 =7
FILE 1
As many as six descriptors may appear with an asterisk, indicating that they are the "major' descriptors. (Themajor descriptors are used as subject index entries for the printed issues of RIE and It IJE.) Searching can berestricted to the major descriptors by the use of the /DE and /DE suffixes as illustrated below or by theuse of the LIMIT command discussed on page 1-213_
(11 1_28)
When the searchable indexes are created, all multiword descriptors are re-indexed under each meaningfulindividual word of the phrase_ In addition, it is possible to differentiate between single-word descriptors andthe same single words extracted from multiword descriptors. The following example with the term "reading ""illustrates the varying degrees of specificity and differences in postings.
.29)
? SELECT READING; SELECT READING/DE; SELECT READING /DEB8 22699 READING9 17104 READING/DE
10 12555' READING /DE'? SELECT READING/DE; SELECT READING DF*
11 2538 READING/DE12 864 READING /DF*
In set 8, the term is retrieved from any suffix-coded field in the basic index. In set 9, the term is retrievedfrom single - -Nord or multiword descriptors, such as "reading," "reading ability," "reading achievement,- "readinginstruction," "critical reading," "supplementary reading materials," etc. However, in set 10 the term is retrievedfrom these descriptors only if they have been asterisked as major. In sets 11 and 12 retrieval is restricted tothe single-word term READING, and in the latter instance the word must have been asterisked as a majordescriptor.
(1) 1.30)
When the multiword.descriptors are re-indexed under the individual words of the phrase, a few common wordswithout subject content are omitted from the index. To search for phrases containing such "stop" words, thedescriptor phrase can be SELECTed as a unit, e.g.,
(ii 1.31)
? SELECT TRANSFER OF TRAINING13 802 TRANSFER OF TRAINING
However, if full-text methods are used, the "of" must be excluded, but counted, e.g.,
? SELECT TRANSF ER( 1W)TRAINING/DE14 802 TRANSFER( 1W)TRAINING DE
The search can be broadened by removing the restriction to the Descriptor ield, e.g_,
? SELECT TRANSFER(1W)TRAINING15 833 TRANSFER( 1W)TRAINING
(11 1.32)
(ft 1.33)
The additional records retrieved by set 15 have the phrase in the Abstract, Identifier, or Title fields.(11 1.34)
FILE 'I
IDENTIFIER( /ID, /ID*, /IF, /IF"). Identifiers are single-word or multiword terms assigned to provideadditional subject indexing beyond descriptor terms. Identifiers are generally free-language terms and tend tobe more specific than the Thesaurus vocabulary, e.g., project names, legislation, geographic locations, andother subject terms which have not yet achieved broad acceptance. As many as two identifiers per record maybe asterisked to designate them as "major.-
(1 1.35)
When identifier phrases are entered into the basic index, they are allowed a maximum of 42 characters tnlength_ To search for lengthy phrases, the searcher can utilize the truncation capability, e_ g.,
? SELECT READING AND MATHEMATICS OBSERVATION SYS?16 4 READING AND MATHEMATICS OBSERV
? TYPE 16
1
EJ149053 SPSOSC53Teacher AptitAes, Knowledge, Attitudes, and Cognir
Predictors of Teaching BehaviorEkstrom, Ruth B. Journal
Win 76
Descriptors: *Teacher Behavior/ *Teacher Attitudes/ *Teaching Styles*Teacher Characteristics/ *Elementary School Teache *Cognitive
Ability/ Teacher Qualifications/ Predictor Variables/ Aptitude/ VerbalAbility
Identifiers= *Beginning Teacher Evaluation Stud APPLE ObservationSystem/ Reading and Mathematics Observation System RAMOS)
1" 1.36)
e Style as
Teacher Education; 27; 329 31
However, since identifiers are also indexed under each meaningful individual word, full-text methods can beused, e.g,,
? SELECT READING(WMATHEMATICSMOBSERVATION/ID17 4 RIADING(WMATHEMATICSMOBSER
(Note that while the command has been processed properly, the resulting t title is shortened.)
(11 1.37)
(II 1_38)
Geopolitical identifiers currently begin with the name of a state (U.S.), province (Canada), or country (allothers) and include the local name as a parenthetical qualifier.
(1) 1.39)
However, prior to January 1976, the area and local names were entered as separate identifiers. Thus, if asearch involves a specific location, both formats can be retrieved by use of the F full-text operator, e.g.,(11 1.40)
SELECT MEMPHIS(F)TENNESSEE/1018 23 MEMPHIS(F)TENNESSEE /ID
? TYPE 181
E1148118 EC090028The Educational, Behavioral, and Psychological Characteristics
Teachers as Motivators for Creative and Less Creative Students
179,
FILE 1Hooks, Mose Yvonne Brooks North Carolina Association for
G i f t e d and T a l e n t e d Q u a r t e r l y J o u r n a l ; 1! 4; 52-60 F 75Descriptors: *General Education/ *Creative Ability/ 'Creativity
Research/ *Teacher Characteristics/ *Student Motivation/ Res rch-Proiects/ Elementary Secondary Education
Identifiers; Tennessee IMemphisl
? TYPE 18/2/2323
ED001246THE MEMPHIS STORY-MOBILIZING A LARGE SCHOOL SYSTEM AND A COMMUNITY
FOR THE SUPERIOR AND TALENTED STUDENT PROJECT,COUNCE, D. SHELBYNorth Central Association of Coll. and Secondary Schools. Chicago,
Ill.; Southern Association of Colleges and Schools. Atlanta, Go.Pubis Claret 65 Note: 68P.EDRS PRICE MF-$0.76 HC -$3.32 PLUS POSTAGEDescriptors: Academic Aspiration/ Curriculum Developmept *GuidanceServices/ *Identification/ Occupational Choice/ Special Programs/
Student Motivation/ *Superior Students/ *Talented StudentsIdentifiers: ATLANTA GEORGIA/ CHICAGO/ ILLINOIS/ MEMPHIS/ STS
PROJECT/ TENNESSEE
In actual searching, it would be best to include the other relevant suffix -coded fields in the entry, e.g.,(11 1.41)
? SELECT MEMPHIS(F)TENNESSEE/ID,AB,TI-.19 F5 MEMPHIS(F)TENNESSEE/ID
The above entry is generally better for geographic area searching than an'unsuffixed entry which would retrievealso from the /NT, /CS, and /SA fields:
(11 1.42)
Personal names may be used as identifiers and generally are entered with surname first. However, forcomprehensive searching, the entered tern should have a format compatible with all fields in which it is likelyto occur. For very well-known names, the surname may be specific enough, e.g., SELECT SHAKESPEARE.For less unique surnames, the initials and/or name should be included in the entry, e.g., SELECT E(W)B(F)WHITE.
(11 1.43)
When the searchable indexes are created, each identifier are indexed as the whole phrase and re-indexed undereach meaningful individual word of the phrase. In order to differentiate between single-word identifiers andthe same single words extracted from multiword identifiers, the /IF and /IF* suffixes can be used Forexample,
? SELECT STANDARDIZATION /ID20 13 STANDARDIZATION /ID
1.44)
retrieves such identifiers as "standardization," "test standardization,- and "international organization forstandardization," while
($ 1.45)
TANDARD I ZAT I ON/ I D*
21 5 STANDARD I ZAT I ON/ I D*
retrieves these same identifiers only if they have been asterisked as major, but
SELECT STANDARDIZATION /IF7 STANDARD I ZAT ION I F
= retrieves only the single-word identifier, and
SELECT STANDARDIZATION/IF*23 2 STANDARDIZATION/ F*
ieves the single-word identifier only if it has been asterisked as major.
(11 1.46)
(11 1.47)
DESCRIPTIVE. NOTE ( /NT). This field contains pagination and other cataloging information to augmentthe document description, such as the circumstances under which the document was prepared (conferencepaper speech, reprint. preprint, ea.), the language of the document, relation to other document% and legibilityOf the document. This field is not used for CUE documents.
(11 1.49)
SPONSORING AGENCY ( /SA). The Sponsoring Agency field is used to list the institution, e.g.. governmentagency, private foundation, etc., other than the corporate author, which supported the work or production ofthe document by providing funds via contract or grant. This field is used only for RIE items. Each sponsoringagency name has been entered in a standardized form and corresponds to a specific eight-character alphanumericcode searchable in the SC= field. When the sponsoring agency is also the corporate author, this field is notuse
In searching the /SA field, fullexample,
(11 1:50)
hods should be used with the most unique words of the name. For
(11 1.51)
ECT CH I LD(W)DEVE LOPMENT/SA24 452 CHILD(W)DEVELOPMENT /,SA
retrieves the records of reports sponsored or cosponsored by the Office,of Child Development DHEW),Washington,
(11 1.52)
Since the Sponsoring Agency field is not used if the report is sponsored by the issuing agency, the searcher islikely to want to retrieve reports sponsored or issued by a particular agency. This is best achieved by usingboth suffixes with the entry, e.g.,
(11 1.53)
181
FILE 1? SELECT CHIL4(W)DEVELOPMENT/C5,5A
25 700 CHILD(W)DEVELOPMENT CS SA
TITLE (/TI). The complete title is entered as it appears on the original document, including alternative title,subtitle, and other associated descriptive matter. Fabricated titles, for documents without an original title, areenClosed in brackets.
(II 1.54)
All meaningful individual wards from the Title are searchable. Any hyphenated words are separated into theircomponent parts and must be searched using full-tc-At methods. Thus, to find titles with the term "non-Englishspeaking," the following command could be entered.
? SELECT NONMENGLISH(W)SPEAKiNG/T126 46 NONMENG1ISH(W)SPEAKING/T1
? TYPE 26/6/1-2
EJ15127.9
The Non qlaLLILmiila Chi I'd in Your Classroom
2
EJ146312
The Puerto Rican Non-En lish-S eakin Child: What Can I Da?'
Note that both the hyphenated and non-hyphenated forms are retrieved.
1.55)
(1 1.56)
LEGISLATIVE AUTHORITY CODE (AC.). For RIE documents resulting from research funded by theOffice of Education or by the National Institute of Education, a two-digit code indicates the legislative authorityunder which the project was funded. The following codes are used in this field:41
CODE PROGRAM AREA
0$ Adult and Vocational Education, PL 88-210
16 Captioned Films for the Deaf, PL 85-905
24 Cooperative Research, PL 89-10, Title 1V
32 Disadvantaged Students Program, PL 89-10, Title I
Handicapped Children and Youth, PL 88-164
48 Language Development, FL 85-864, Title VI
52 Library Research anil Development, PI- 89-320, Title HI, Part B
56 New Educational Media, FL 85-864, Title VII, Part A
New Educational Media, PL 85-864, Title VII, Part B
182
CO 1.57)
FILE 1
CODE PROGRAM AREA
,72 Research in Foreign Countries, FL 83-480
80 State Educational Agencies Experiment Activities, PL 89-10, Title V, Section 505
88 Supplementary Centers and Services, PL 88-10, Title Ill
95 Other Office of Education and National Institute of Education Programs (Projects(Or which the legislative authority cannot be determined are included here.)
The legislative authority code appears after the ED number, as illustrated in the following example.(11 58
? SELECT AC=4027 AC=40
TYPE 27/2/22
ED129.532 40 RC009498Outdoor Education for Emotionally Handicapped Students m Central
New York. Final Report, September 1975. A Report, ori Proiect 75-940 forHandicapped Children Under Title VI-B (August 1, 1974-Jul 31, 1975).Miller, JamesPubl. Dote: Sep 75 Note: 70p,EDRS Price MF-$0.83 HC -$3.50 Plus Postage.Descriptors: Communication Skills/ Comparative Analysis/ Educational
Objectives/ *Effective Teaching/ *Emotionally Disturbed/ EnvironmentalEducation/ Formative Evaluation / Motor Development/ *Outdoor Education/ Participant Involvement/ Pilot Projects/ *Program Evaluation/ SelfConcept/ Social Development/ Special Education/ *Student Attitudes/*Student Behavior/ Student Development/ Teacher Evaluation
identifiers: Elementary Secondary Education Act Title VI B/ ESEATitle VI 8/ *New York (Central)
AUTHOR (AU= ). Up to two individual authors per document may be listed in an online record. If thereare more than two authors, only the first author is listed followed by "And Others." Author names generallyappear as published in the source document. The AU. index has been EXPANDed to illustrate the formatof the entries.
EXPAND AU=MOORE,--JOHNRef Index-term Type Items RTEl AJ=MOORE. JOHN 2E2 AU=MOORE, JOHN N.----- 3E3 AU=MOORE, JOHNE4 AU=MOORE, JOHN R. 1
ES AU=MOORE, JOHN ROBERTE6 -AU=MOORE, JOHN 8E7 AU=MOORE, JOHN 7
(ii 1.59)
183
ES AU=MOOR E J OHNNA
E9 AUmMOORE, JOSEPH A..:-E10 AUmMOORE, JOSEPH 5
AUmMOORE JOSEPH T.E12 AU-MOOPE, JOSIAHE13 AUmMOORE, JULIE L. 2
E14 AU,MOORE, JUNE 6
E15 AUmMOORE, JUSTIN REI6 AUmMOORE, K. E. 1
E17 AUmMOORE, K. L.
EIS AUmMOORE, KAPE- I
E19 AUmMOORE, KATHRYN M. 4
-more-
FILE 1
Author names may include full name or initials, or some combination of name(s) and initials with spacingand punctuation. Variations exist, depending on the source document. If the individual edited, rather thanwrote, the document, "ED." is appended to the name as in the entry at E7. EXPANDing is recommendedprior to SELECTing an author name.
(1; 1.60)
SELECT E6, E7
28 ,15 E6. E7
E 6 : AU- MOOR E JOHN W.
CLEARINGHOUSE CODE (CH=). Currently there are 16 clearinghouses located at various universitiesand professional organizations throughout the United States which are responsible for acquiring, indexing andabstracting the reports and journal articles relevant to _their respective subject area specialty.
(11 1.61)
Each clearinghouse assigns a temporary accession number to each document it inputs. The clearinghouse- accession numbers searchable in the CH= field consist of a two-letter clearinghouse code followed by a six-digit
number. e.g., CH=SE506669.-(11 1.62)
All clearinghouse accession numbers have been cascaded to the two-letter clearinghouse code, e.g., SELECTCFI SE retrieves all documents-indexed by the Science, Mathematics, and Environmental Education Clearinghouse:The clearinghouse code can be used as a broad subject area restriction in a search strategy as illustrated below.
(11 1.63)
? SELECT SECONDARY EDUCATION29 11293 SECONDARY EDUCATION (EDUCATIO
? SELECT SECONDARY(W)EDUCATION(C)CH=SE30 3153 SECONDARYMEDUCATION(C)CHmSE
In set 30 the C`fo-IrteXt operator has been used to restrict the term "secondary education" to documentsindexed by the Science, Mathematics, and Environmental Education Clearinghouse.
1.64)
Over the timespan of the database, there have been some changes in the number of clearinghouses as well asin their codes. The current clearinghouse codes and previous codes are listed in Table 1, To search in a
FILE 1
subject area where the clearinghouse was. discontinued and merged with another, SELECT the old code andthe current code. For example, for a broad search on Library and Information Sciences,
(- 1.65)
L ECT CH= LI , SELECT C`! =IR,. COn181PE 31 R 3231 6228 CH= Li32 8506 CH = IR
33 14734 31 OR 37
-TABLE 1. Clearinghouse Codes (CH=)
AA - ERIC Facility Contr. nes discontinued with iarclt 1973 R1E)
AC Adult Education (merged into CE in 1973)
Al Linguistics (merged into FL in 1971)
CE - Career Education
CG - Counseling and Personnel Ser lees
CS - Reading and Communication Skills
EA Educational Mauagcrnent
EC - Handicapped and Gifted Children
EF - Educationul Facilities (merged into EA in 1970)
EMEducational Media and Technology (merged into IR In 1974)
FL -Language d Linguistics
IIE - Higher Education
IR - tntormation Resources
JC - Junior C011eges -,
LI - Library and Informat
PS - Early Childhood Educalioin
RC. - Rural Education and Small Schools
-RE -Readi (merged into CS in 1972)
SE -Science, Mathematics, and Ensironmental Education
So - Social Science/Social Studies
n (merged into IR in 1974)
-Teacher Education
ching of English (merged into CS ul 1972)
Tests. ,Measurement, and Evaluation..
(tnergcd into CI
FIRE )
CONTRACT/GRANT:NUMBER (CN-). If the contract or grant numbers assigned by sponsoring agenciesarc plainly ,apparent on the document, they are included in this field. Prior to July 1976, only U.S. Office ofEducation (OE) or National Institute of Education (NIE) grant and contract numbers were indexed in thisfield.
(C.1.66)
,atria numbers generally include to alphabetic prefix followed by a variable number of digits.. Some typical :Tact and grant numbers are
0E-0-73:7094
(_D-C13-02(c4)
N I E-C400.75-0015
N61339-73 C'.0'0.7
All contract and grant numbers arc cascaded to the initial alphabetic prefix. Thus SELECT CNOCD includesCN:- OCD-CII-02(C4) and all other contract numbers beginning with "OCD
used (or CITE items_
(r 1.68)
DOCUMENT TYPE (DT, ). Since September 1974. a single- character alphabetic code has been used in thisfield in indicate the form of the_RIE document The codes and their corresponding publication types arc li,.tedin Table 2.
,70)
Itt a document type code automatically restricts retrievel to on enteredn the ERIC database itnee September 1974:
r bibliographies on the "gifted.:" the following command could' be
CT DT.t (CiGIFTED34 30 Dr=tiC)GIFT
FILE 1
TABLE 2. Document Type Codes (DT --=.)
Audio Visual/Nonprint Media; Audiovisual Aids; Films; Tape Recordings; PhonotapeRecordings; Computer Programs;' etc.
Books; Monographs, Textbooks; Programmed T s; etc (not otherwise classifiable)
Curriculum Guides; Teacher-Developed Materials; Laboratory Manuals
D Directories; Membership Lists; Tables of Organization; Reference Works Dealing withOrganizations/Institutions; etc.
Guides; Teaching Guides; Resource Guides; Study Guides; Administrative Guides;Leaders Guides; Manuals; Training Manuals
Legislation, Legislative Hearings, Legislative Reports, Congressional Documents.(Include both Federal and State levels; include National Comminsions).
Journal Articles; Serials; Periodicals; Bulletin Newsletters; Newspapers; etc_
Program/Project Descriptions
Bibliographies; Annotated Bibliographies; Book Catalogs; Abstracts; LiteratureReviews; Literature Searches/Guides; Book Lists; Book Reviews; Library Guides;Indexes (Locators); State-of-the-Art Reviews
Maps; Atlases; Gazetteers
Numerical and Statistical Tables; Quantitative Data and Analyses
Other
Proceedings; Conference Records/Minutes (entire)
Questionnaires; Tests; Measurement Devices; Evaluation Devices
Reports (Research and Technical)
Speeches; Conference ReportS; "Papers presented at. Verbal Presentations
Theses; Dissertations
Dictionaries; Vocabularies; Glossaries; Thesauri
Annual Reports; Yearbooks
GROUP CODE (GC= ). Specific Thesaurus terms appropriate to very broad subject areas have been groupedtogether in the Group Code field. Each code is a three-digit number and represents a group of related subjectterms, e.g., GC=060 for Behavior terms.
FILE 1
Group codes cannot be SELECTed directly. The EXPAND command can be used with a group code inparentheses to see the terms included in the group and their associated postings. Individual terms, or all termsIncluded in the group may be SELECTed by their R reference numbers as illustrated below.
(II 1.74)
EXPAND (GC=3601Ref
RI
R2
R3
Index-term 'Type-GC=360 (OPPORTUNITIES)AFFIRMATIVE ACTIONCAREER OPPORTUNITIES - - - =D
Items
293
1037
RT
14
22
JO.
R4 COMMUNITY BENEFITS --. D 87 7
R5 CULTURAL OPPORTUNITIES D 55 5
Re ECONOMIC OPPORTUNITIES D 165 3
R7 EDUCATIONAL- OPPORTUNITIES 0 1770 9
RS EMPLOYMENT OPPORTUNITIESD 2450 29R9 HOUSING OPPORTUNITIES D 69 4'
RIO OPPORTUNITIES D 10742 10'RII RESEARCH APPRENTICESH1f5D 23 2
R12 RESEARCH OPPORTUNITIES D 254 2
R13 SOCIAL OPPORTUNITIES D 80 8
RI4 TEACHING BENEFITS D 58 2
R15 YOUTH OPPORTUNITIES D 160 4
? SELECT R1-R1535 11035 RI.R15
R1: GC=360 (OPPORTUNITIES)
The 52 broad descriptor groups and their codes are listed with their respective scope notes in the Thesaurusof ERIC Descriptors.
(11 1_75)
ISSUE (IS.). This field is used to indicate the monthly issue of RIE in which the document is announced,e.g., IS ItIEAPR77 for all documents announced in the April 1977 issue of RIE. For documents announcedin CUE, only the year of announcement is indicated, e.g., IS=CIJE1977.
(11 1.76)
JOURNAL NAME (IO= ). Currently_ , full journal titles, with the exception of initial articles, are enteredthis fielc:. Until 1971, periodical titles were abbreviated according to ANSI Standard Z39,5-1963 "Periodical
Title Abbreviations.- For comprehensive searching on a specific journal title, both forms should be SELECTed,
? SELECT JO=PUBLIC OPIN QUART, -SELECT JO=PUBL IC OPINION QUARTERLY36 24 JO=PUElt IC OPLN QUART37 42 JO=PUBL IC OPINION QUARTERLY_
COMBINE 36 or 3738 66 36 OR 37
(5 1.77)
field may be used for _RIE documents,
38
FILE 1
BUREAU/PROJECT NUMBER (PN--=). This field is used to list the alphanumeric code assigned by thesponsoring agency to the project for which the document was produced. Currently only OE or NIE projectnumbers are indexed.
($ 1.79)
Project numbers generally include an alphabetic prefix followed by a variable number of digits and/or additionalletters. Some typical project numbers are:
1.80)
BR-H-127245B
CG8119A/5
1,0008.1A
V-361057L
All project numbers are cascaded to the initial alphabetic prefix. Thus SELE PN=L includes PN=L0008JAand all other project numbers beginning with L.
Thi fiel is no d for CUE items.
(r. 1.81)
(II 1.82)
REPORT NUMBER (RN= ). This field is used to index the unique identifying nunber(s) assigned to the
(11 1.83)
publication by the organization(s) producing or issuing the document.
Report numbers generally include an alphabetic prefix followed by a variable number of digits and/or additionalletters. Some typical report numbers are:
($ 1.84)
AAAS-MISC-PUB-76-2
DA-PAM-550-152
RX74.15-HEW
T-75.-196-G
All report numbers are cascaded to the initial alphabetic prefix. Thus SELECT RN=RX includesRN = RX74-15-HEW and all other report numbers beginning with RX. Some entry variations in format ofreport numbers may exist, so it is best to EXPAND in the RN= index to locate a specific report by its number.(ii 1.85)
This field is not used for CUE items.
(11 L86),
CORPORATE SOURCE CODE or SPONSORING AGENCY CODE (SC=). An eight-charactr alphanumericcode assigned by ERIC to each originating institution and each sponsoring agency. An alphabetic directoryof the institutions, sponsorang agencies and their codes is available (see the Search Aids section). Codes includethree letters followed by five digits, c.g SC= AHP00825 for the Alabama State Dept. of Education. Montgomery.or.SC=BBB08152 for the Office of Education (DREW). Washington. D.C., Right to Read Program.(II 1.87)
This field is not used for CUE documents.
FILE 1
(Ii 1.88)
UPDATE (U0,-). Search results can be restricted to a given update of the database, e.g., SELECT UD=7705.In addition to the individual update numbers, UD=9999 is always identical to the latest update in the database.A SELECTed update set can be COMBINEd with a subject set to retrieve citations in that update of ERIC.
(Ii 1.89)
YEAR (YR= ). This field is used to indicate the last two digits of the year the document was announced inRIE or CUE, However, to restrict retrieval to a range of years, it is More convenient to use the LIMITcommand as described in the following section.
LIMITING
(1( 1.90)
The LIMIT command is used to reduce a set according to some criterion applicable to the database. Generallythis command should. be entered after SELECTing the appropriate set. Sets SELECTed in the ERIC databasemay be LIMITed by subfile, by accession number within each subfile, by document availability, and by major9e minor descriptor or identifier.
(li 1.91)
ED or EJ Subfillo. A SELECTed Set can be restricted to the report literature subfile (RIEliterature subfile (CUE) by use of the /ED or /EJ suffixes respectively as illustrated below.
? SELECT SCIENCE(W)EDUCATION
39 11003 SCIENCE(W)EDUCAT NLIMIT 9/E11; LIMIT 39/E1
40 3649 39/ED41 7354 39/E1
the periodical
(11 1.92)
Accession Number. The DIALOG accession number which appears in each printed record is an eight-characternumber Which includes the ED or El prefix followed by six digits and corresponds to the accession numberwhich appears on each record in the printed RIE or CBE To LIMIT by accession number range, the prefixis dropped and the appropriate suffix is used with the accession number range.
(11 1.93)a
The annual accession number ranges_ that may be used with the LIMIT command for each subfile are listedbelow. This table is also available online in response to 7YEARS1.
YEAR ED NUMBERS EJ NUMBERS
PRE-66 002747-003960/ED-1966- 010000-010093/ED19671968196919701971
19721973
1974
010094-012348/ED012349-021!51/ED021152-031604/ED031605-042060/ED042061-054390/ED054391.066620/ED066621.080787/ED080788-095253/ED
000001-01 1 707/EJ
011708-027599/EJ027600-045271/EJ045272-062751/EJ
062752-082164/E1082165-101872/EJ
(11 1.94)
1975 095254-110594/ED 101873-121926/EJ1976 110595-127413/ED 121927-142252/EJ1977 127414- 142253-
To restrict a set to time period, the appropriate DIALOG accession number range for each subfile isused and the resulting seits are COMBINEd using the Boolean OR. For example, to LIMIT set 39 to thosedocuments announced id ERIC from 1974 to the present:
(11 1.95)
? LIMIT 39/080788-999999/ED; LIMIT 39/082165-999999/EJ42 2335 39/08078B-999999/ED43 5576 39/082165-999999/E1
? COMBINE 42 OR 4344 7911 42 OR 43
In LIMITing by accession number range, necessary to have all of the digits filled with numbers. Toretrieve citations including the very newest entries, it is possible to "fill" the upper range with nines as shownin the previous example.
(11 1.96)
Document Availokaty: A SELECTed set can be restricted to those documents for which full copies areavailable or not available trim the ERIC Document Reproduction Service (EDRS) by the use of the suffixs/AVAIL and /IJNAVAIL respec!ively.
1.97)
? LIMIT 39/AVAIL:- Li MIT 39/UNAVAIL45 2586 39/AVAIL46 8417 39/LINAVAIL
Since copies of documents announced in the CUE subfile are not available from EDRS, the available documents-are always from the RIE (or ED) subfile.
(V 1.98)
Major or Minor Descriptor or Identifier. The asterisked descriptors and identifiers (which are also thesubject ,index ergries in the printed RIE and CUE) are considered to be major terms. A search may be LIMITedto major terms only by using /MAJ The /MIN suffix LIMITs the set to all other occurrences of the term.
(V 1.99)
The /MAJ and /MIN LIMIT'suffixes are properly applied to sets formed from terms SELECTed in theDescriptor and/or Identifier fieleN.
(11 1.100)
However, it is gdierally preferable to restrict a search to major descriptor or identifier by using the /DE* andD fields, e.g.,
5'fLECT CIENCEMEPLICATION/0E*,I0*47 2825 5CIENCEMEDUCATION/DE*,10*
(11 1.101)
Multiple Limits. Multiple LIMITS can be used if they are entered in the acceptable order, namely theaccession number range precedes the subfile. which precedes the rnajor/rninor term, which precedes the
FILE 1
ailability/nonavailabilityspecification. For example, to apply all of the LIMIT restrictions to the set on.Science Education, the following entry can be used:
47.1.102)
LIMIT 39/080788-999999/ED/MAJ/AVAIL48 480 39/080788. 999999 /ED /MA) /AVAIL
However, the above entry is redundant. Since the availability LIMIT restricts the set to only ED documentsit is not necessary to include the ED suffix in the entered command-, e.g..
? LIMIT 391080788-999999/MAJ(AVAlt49 480 39/080788-999999/MAJ/AVAIL
The gen
(11 1.103)
at order listed above must be followed even when only some of the LIMIT specifications are used.
(! 1.104)
LIMIT 9/080788-999999/ED/MA150 647 39/080788-999999/ED/MAJ
IT 9/MAJ/AvAIL51 839 39/MA1/AVAIL
ITALL. The LIMITALL command can be used in two ways: to LIMIT to an accession number rangeto restrict the search to ED or El documents. The LIMITALL command can be used with the /ED or
/EJ suffix to restrict all subsequent sets to the subfile specified by the suffix, until the command is canceledwith the-LIMITALL/ALL command, or a new LIMITALL command, e.g.,
(f 1.105)
2 LIMITALL/ED'LIMIT ALL ALL/ED
2 SELECT-SCIENCE(W)EDUCATION52 3649 SCIENCE(W)EDUCATION
The accession, number range can be entered with the LIMITALL command to restrict all subsequent sets tothe time period specified by the accession numbers within the appropriate subfile, until the command is canceledwith'the LIMITALL/ALL command, or a new LIMITALL command, e.g.,
IIMITALL/080788-9999994EDLIMIT ALL AIL/080788-999999AED
? SELECT SCIENCEMEDUCAT'ION53 2335 SCIENCE(W)EDUCATION
IAMITALL/ALLLIMIT ALL ALL/ALL
(11 1.106)
FILE
FORMAT OPTIONS
If a format is not specified in TYPE, DISPLAY, or PRINT commands, format 2 is assumed, The records aredisplayed-with-the-highest accession number (newest record) fiTit, within each subfile of File 1. All COE (El)records precede all RIE (ED) records.
(IT 1.107)
FORMAT 1: -DIALOG Accession Number
.108)
EJ034210 EJ025800 EJ021833 EJ013689 EJ009626 EJ008710 ED132588ED131621 ED131416 ED127236 ED126505 ED099878 ED099001 ED098968
FORMAT 2: Complete Record Except Abstract(11 1.109)
CUE
RIE
EJ150129 CS710249The Cloze Procedure; Con It Be Used with Li r_ture?Dupuis. Mary M. Reading Improvement; 13; 199--203 WDescriptors: *Cloze Procedure/ *Readability/ *Literature/ *Reading'
Comprehension/ *Reading Ability/ *Predictive Validity/ ReadingResearch/ Secondary Education/ Grade 10
ED131416: CS003014Reading Guidance in a Media Age.Palette, Nancy; Hamlin, MarjoriePubl. Date: 75 Nate; 260pAvailable from.: Scarecrow Press, Inc, 52 Liberty St., P.O. Box 656,tuchen, New Jersey 08840 ($10.00 cloth)Document Not Available from EDRS.Descriptors; Audiovisual Aids/ *Childrens Books/. *Childrens!lercirure/ -*Early Reading/ Elementary Education/ Elementary School
libraries/ *Literacy/ *Mass Media/ Reading Development/ *ReadingInstruction/ Reading Materials Teaching Techniques
FORMAT 3c Bibliographic Citation
CUE
EJ15012.9
The Cloze Procedure; Con It Used with-Literature?Dupuis Mary M.Reading Improvement,W 76
(T 1.110)
RIE
ED131416#
Reading Guidance in a Media Age.Palette, Nancy;' Hamlin, Matiorie260p. 75Scarecrow Press, 'lac., 52 Liberty SI: P.O. Boa 656_ Metuchen, New
Jersey 08840 ($10.00 cloth)Document Not. Available from MRS,-
FORMAT 4: Abstract
C1JE
RI
EJ150129Reveals that the cloie 'and comprehension -test scores correlated at
significant levels, thusdoze procedure. (RB)
1
supporting the predictive validity of the
ED13141#
Chapters An this voAerme discuss the problems that teacher parents,and librarians face in encouraging elementary school. children to readand suggest ways to improve student motivation and reading inthis age of mass media. "Sparking the Fuse for Literacy" defines thecomplex term "11tercicy,® and examines the effects of the media onreading. "Home his Where the Sgart As" AAscuSses parental roles inreading development, while "Library Lures to Literacy" suggestsactivities to stimulate student interest- Other techniques- designed tonliven the reading process ore 'provided Ari "Independent-- StudyPrograms imiLiterciture," From Page to Stage -. -Via Drama and Puppetry,"and 7Sharing Literature: The Audiovisual Experience," "Sound; Visuals,Acton " explores.the development of g student production ce.riter usingvideo and audio eauipment.=Conference pli'-rning and introducArig authorsto children through audi6Visuat _medic, are discussed in "Students,Teachers; LiOrarions and Authors Get:Tagether-." In addition, eachchapter contains o brief bibliography of children's books. (KS)
FORMAT 5: Fu
CUE
Record
EJ150129 CS710249The Cloze Procedure: Can It Be Used with Literature?Dupuis, Mary M. Reading Improvement; 13; 4;,199-203Descriptors: *Cloze Procedure! *Readability/ *Literature!
Comprehension/ *Reading Ability /., *Predictive ValidResearch/ Secondary Education/ Grade 10
Reveals that the-cloze and comprehension test scores.aarrignificant levels, thus supporting the -predictive validi_
Ooze procedure. (RB)
a
I a a
(1 1.112)
76*ReadingReading--
fated atV of the
FILE I
1:137416: CSO030i4keacing Guiaoace in a Media Agevoiette. Norc. , Har. marlorie
Dote 75 Nolo 260pAxo.lable frotr Scarecrow Pais
. 521bervSt, PG 130.x 6.56-Me'toen Nev Jersey 08840 i$10 00Dt Nci Available fro EDRSDesoaotors Aualovis-..al Aids *Childrens heoxs: *ChildtensL.teftwr,' 'Lakly Reed 7a,- Ete-cItary Educe!ian Elementary SchoolItorates eracy- -Mass Media Reading Development/ *kcodingReodIng Mater,ais TeaLLinq TechniquesCaotefs ia tt s YOlum r. discuss the problems tnat teachets.-parents.a-o f!otterans face In encouraging elementary school Children to read-suggest ways to improve studenr
motivation one reading obiltry inage- di mass media. -"Sparking the Fuse fo, Literacy defines ts-er- 'Hiterocy.- and examines the effects 0 thefead i'a 'Ham,- is Where the Start Is" discusses parental voles inread!-4 deveIonrwent. while -Library Lures to Literacy>---suggests, acyx.iies to stimulate staent interest.
Other techniquel-designed toenliven the reading process are provided in ",..11*(dependent StudyPrograms in Literature." n.From Page to Stage--Vama and Puppetry."and "Sharing Literature The Audiovisual -Expv.,--4nce." "Sound, VisualsAction: esplores the development of a sit:Oent production center using,deo Or,d audio eguipmer-d. Conference-aXnning and introducing outhorsL.xtilaien 'nrough audiovisual,..mia are discussed in "Students,Teache's tibioiions and Auth9ki Get Together.- In addition, eachcptet COtO n5 a brief bibli taphy of children's books. (KS)
FORMAT 6: Title
CUE
'-EJ151.11c,Ane ocedure' Can ft Be Used with Liter ure?
Ebt:i(4ft
Wead:cng.Guidonce in a.Medio Age.
Wrect Access. Any le,..ord in the file may be TYPEd or DISPLAYed direetly m any format using the)IALOG AccFssion Number/Format in the TYPE or DISPLAY command, e.g.,
i TYPf E 1150179
EJ150129 -
Reveals tv:Ir the clo2e and Comprehension test scores correlated at5,wl Iekt levels
;thus supportfng the predictive validity of the';0cize procedure. IRV)
Ins
FILE
SEARCH AIDS
Thesaurus of ERIC DescriptorsThe complete ERIC socabulary of descriptors is listed alphabetically. hierarchically. and in a rotateddisplay
updated anually. $7-.95Asailable from Macmillan. Inc
(` 1-119)
Descriptor and Id,nrifier Usage ReportLists each ERIC i'escriptor and identifier and the.number of times it has heci used to date1966-77 Accumulated totals, 53.50Available from the ERIC Clearinghouse on Science, Mathematics; and Ens ronmental Education
1.120)
Institutional Source DirectoryAn alphabetical Listing of the names of all institutions by which documents in the ERIC system havebeen inddxed in the Corporate Source Name and Sponsoring Agency Name fields of the citation, togetherwith the corresponding alphanumeric code which appears in either the Corporate Source or SponsoringAgency Code field of each recordPublished .sermanually. $15 domestic, $18 foreignAnnual subscription. $25 domestic, 530 foreign.Available from the ERIC Processing and Reference Facility,
(c 1.121)
Directory of ERIC Microfiche CollectionsA listing of the over 600 libraries and information centers that have a standing order to the ERICmicrofiche collection. The directory gives collection scope/size, equipment and services and accessibilitydata, as well as the name and phone number of a contact person at each collection. Many of thesefacilities are open to the general public, and will duplicate microfiche and/or hardcopy.Published annually; free on requestAvailable from the ERIC Processing and Reference Facility.
ff 1.122
Survey, of ERIC Data Rase Search ServicesA listing of the sites which provide search services on a regular basisirrespective of whether the serviceis available only to a circumscribed community or . all users without restriction.Published irregularly, free on requestAVailable from the ERIC Processing and Reference Facility-.
(C 1.123)
Interchange NewsletterThe newsletter used by the ERIC Processing and Reference Facility to communicate with ERIC usersabout changes in the system, search tips, and periodically held ERIC users meetings.Published irregularly; free of charge to interested organizations
(1- 1.124)DOCUMENT RETRIEVAL
Copies of non-copyrighted documents (and any others for which reproduction permission has been obtained)announced in the RIE subfile of the ERIC database are available from the ERIC Document ReproductionService (ERRS), P.O. Box 190, Arlington, Virginia 22210.
(f 1.126)
Copies may also be available from many of the facilities listed in the Directory of ERIC Collections. describedin the Search Aids section.
1.127)
INDEX
Answer sets to ONTAP searcl-, topics 49 -51, 61
BEGIN command 40Boolean operators 15-17, 24-25Briefsearch 5, 10, 11, 19, 10-21,'33,40, 45
examples of, in self-improvement exercises 66-67, 76 90, 94, 98,102, 106, 112, 119, 124, 133
Building block approach to search strategy 17-19examples of, in Search Save formulations 67 -68, 143-144, 147-17Dexamples of, in self-improvement exercises 57, 68-69, 81-82, 90, 98,
102-103, 113-114, 124. 128, 133, 134-135
Citation pearl growing approach to search strategy 8 -21, 30examples of, in self-improvement exercises 117-119
Compiling the search terms 6, 9, 29-31Computer working meMory 18, 37-43Concept groups- See Facet analysis.Conceptualizing the .search as input to the retrieval system 7, 33, 35 -43Controlled term searching 6, 9, 29, 36-37
examples of, in Search Save formulations 163-164examples of, in self-improvement exercises 56-57, 73, 82-83, 98, 102,
119, 139Cost as a search objective 11, 45 47-48, 75-77, 136-137
in self-improvement exercises 75, 93, 135
Data base selection. See Identifying relevant data bases.Defining the information -need and search objective 5, 9-11Descriptor 6
field of ERIC record 29, 36function of 11
"Difficult" ERIC ONTAP search topics, list of 84-
ERIC data base, description of 171-1S6 (appendix).ERIC ONTAP data base, description of 49-53, 58-62/EVAL command 52, 61-62Evaluating final results 7, 47-48Evaluating preliminary' results 7, 34, 45EXPAND command 30, 39, 69
Facet analysis 6, 7, 9, 15-17, 18, 56, 105, 106-107; 131, 135-136 ,
Formulating basic search logic 6, 15-29,Free text searching 6, 9, 29-30, 35
examples of, in Search. Save formulations 67-68, 143-144, 147-162, 165 -170examples of in self-improvement exercises 57, 68-70, 73, 76, 82-83, 90;
94, 98, 102-103, 107, 112-114. 119, 124, 128, 132, 138-140
High precision as a search objective 5, 10311, 47-48in Search Save formulations 163-164in self-improveMent exercises 65, 79, 118, 135
High recall as a search objective 5, 10, 36, 46, 47in Search Save formulations 67-68, 143-144, 147-162, 165-170in self-improvement exercises 55, 65, 75, 79, 93, 105, 111, 123
Identifier 6
field of ERIC record 29, 36function of 11
Identifying relevant data bases 5, 13Interviewing the information seeker 9
LIMIT command 40-41, 80, 111, 123LIMITALL command 23, 41, 123, 137Lowest postings facet first-approach to search strategy 16
examples of, in self-imp: ,,,ment exercises 90-91
"Medium" ERIC ONTAP sear::''; topics, list of 71
Most specific facet first approach to search strategy 25-26mples or, in self-improvement exercises 69-70, 73-74, 93-94 112, 132
Online alphabetic word list 6, 30Online ERIC Thesaurus 6, 30-31Online searching, stops and functions of 5 -7, 9-48, 53ONTAI' data base, description of 49-53, 58-Ordering output 6, 9, 33-34Output
formatlimiting If 6. 7, 9, 23, 34, 86printing of 6, 118size of 16, '3, 25-26,
Precision, definition of 10, 47-48. See also High precision.
Recall, definition of 10, 47. See also High recall.Relevance of citations of ONTAP answer sets 47. 50-51EatELfposaurus 30, 55
Search objectives 5, 9-11, 47, 50, 51Search Saves
common search facets/Search Save formulations 143-170. (appendix)example of 67-68exchanging of, across DIALOG passwords 145to bypass computer memory_ limitations 41 -43
Search strategy, approaches to 6, 9, 17-27, 51Search strategy, formulation of. See Formulating basic search logic."Simple" ERIC ONTAP search topics, list of ..61
Successive fra:ctions approach,to search strategy 22-25examples of, in self-improvement exercises 82-83, 106-107, 127-128, 132
Thesaurus of ERIC Descriptors 29, 30, 49, 55, 111Thesaurus of 30, 55, 111Truncation 33, 36
Word proximity 11, 33-35,
top related